[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/mlp/ - Pony


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: 100.jpg (2.09 MB, 900x1387)
2.09 MB
2.09 MB JPG
Welcome to Chatbot AI General #100, the thread for discussing and improving AI pony chatbots.

▶ MLP Bots
https://mlpchag.neocities.org
Spreadsheet (CAI bots + Old repository):
https://docs.google.com/spreadsheets/d/1J7BeqJVDS51cXF8Pgm2YZaFq-Z6ykSJT
CAI bots converted to Tavern: https://files.catbox.moe/ckurq1.zip
Expression packs: https://rentry.org/ChagExpressions
!!!GALLERY!!!: https://drive.google.com/drive/u/2/folders/1Ao-h5HFGMPllSrzSBKM_BvGSiU9f0c2U

▶ How do I start?
1) Select a Frontend
2) Select an AI model
3) Select Jailbreak
4) Select bots
5) Lovemaking with AI mares!

Starting in this hobby can be confusing and difficult. If it’s your first time and you’re lost,
▶ ASK THE THREAD! ◀

Novice-to-advanced guide: https://rentry.org/onrms

▶ SillyTavern (preferred frontend)
https://github.com/SillyTavern/SillyTavern
On Android: https://rentry.org/STAI-Termux
App that voices pony responses in ST: https://drive.google.com/drive/folders/16Ss26VBmgzcSuTGzhaHqRuyVRceTf-YB

▶ More frontends:
Risu: https://risuai.xyz
Agnai: https://agnai.chat

▶ Locals
https://rentry.org/lunarmodelexperiments
>>>/g/lmg/
Mistral Nemo base model fine-tuned on fimfics: https://huggingface.co/Ada321/Nemo_Pony_2/tree/main

▶ Jailbreaks
MLP JB: https://rentry.org/znon7vxe
More JB and guides: https://rentry.org/jb-listing
Hypebots for Tavern: https://rentry.org/pn3hb

▶ Botmaking
Editors: https://agnai.chat/editor
Guides: https://rentry.org/meta_botmaking_list
Advanced: https://rentry.org/AdvancedCardWritingTricks

▶ /chag/ dustbin of history
https://rentry.org/ChagArchive

Previous thread: >>41638289

▶ Current theme: Neet/Missfit

News:
- Pixtral large out https://mistral.ai/news/pixtral-large/
- Gemini-exp-1114 is now available via the API. Free on Openrouter too.
- Anthropic retire Claude v1 models https://docs.anthropic.com/en/docs/resources/model-deprecations
- 3.5 Haiku out https://www.anthropic.com/claude/haiku
>>
File: Anchor.png (552 KB, 1201x1200)
552 KB
552 KB PNG
Anchor for bots, lorebooks, scenarios.
Previous >>41638292
>>
File: TAnchor.png (287 KB, 973x756)
287 KB
287 KB PNG
Anchor for technical stuffs (Proxies, Updates, Models etc.)
>>
File: ABegging.jpg (96 KB, 559x537)
96 KB
96 KB JPG
Anchor for asking for bots, lorebooks, scenarios etc
>>
File: Screenshots.png (464 KB, 1000x563)
464 KB
464 KB PNG
Bots corner:

Neet Twilight Sparkle >>41641265 https://mlpchag.neocities.org/view?card=Anonymous/Twilight%20Sparkle%20Neet.png
Zabani >>41642592 https://mlpchag.neocities.org/view?card=littlewitcher/Zabani.png
Zephyr Breeze >>41644461 https://mlpchag.neocities.org/view?card=Snowfilly/Zephyr%20Breeze.png

Art corner:
>>41649095

Screenshots corner:

Partial Alicornification Woes >>41638893
Pantsed >>41639299
Unhinged >>41640240
Neet Twilight Sparkle >>41640844
Tactical >>41645828
>>
File: Screenshots2.jpg (1.1 MB, 1748x1181)
1.1 MB
1.1 MB JPG
Screenshot corner 2:

Daddy >>41647094
[Activity you mentioned] >>41647768
Rarity loves you >>41650284
New Gemini Test >>41650732
Horsepussy needs >>41656475
Hatefucking Rarity >>41657080
CYOA preset >>41657480
>>
File: Neet.png (132 KB, 1259x1147)
132 KB
132 KB PNG
Anchor for the Neet/Misfit event!
Everything goes as long as it’s on theme.

Post bots, post logs, post lorebooks...

End of event: 11/27

>Neet Twilight Sparkle
https://mlpchag.neocities.org/view?card=Anonymous/Twilight%20Sparkle%20Neet.png
>Zephyr Breeze
https://mlpchag.neocities.org/view?card=Snowfilly/Zephyr%20Breeze.png
>>
>>41658078
Reposting
rentry.org/CharacterProvider-CYOARPG
Fun preset.
>>
File: making a mare fortress.png (583 KB, 804x762)
583 KB
583 KB PNG
>>41658114
can confirm it works real nice combined with other cards like Pony Colony Simulator
>>
LOVEMAKING WITH AI MARES!
>>
File: Based.png (5 KB, 500x500)
5 KB
5 KB PNG
>>41658161
Based * 100
>>
I believe in mysterymare
>>
File: Princess Cadance1.png (162 KB, 512x768)
162 KB
162 KB PNG
Happy 100th thread /CHAG/. Have some AI mares.

An update of the Candance card I released for Valentines Day, with rewritten definitions and greetings, plus some new ones.

Cadance: https://files.catbox.moe/z13v5f.png
- Updated greeting: A tired overworked Cadance greets you in her throne room.
- Updated greeting: Cadance breaks into your house (again) to review the results of her matchmaking survey.
- Updated greeting: Cadance finds you all alone on Hearts and Hooves Day.
- New greeting: Comrade Cadance seizes the means of affection.
- New greeting: Casual chit chat with the pretty pink pony princess at a pizzeria.

New mares:

Sassy Saddles: https://files.catbox.moe/otg5zv.png
- Sassy greets the newest customer at the Canterlot Carousel.
- You’re Sassy’s new assistant.

Tree Hugger: https://files.catbox.moe/tlm65m.png
- Hanging out at Tree Hugger’s cottage.
- Hiking with Tree Hugger

Songbird Serenade: https://files.catbox.moe/wtkmmg.png
- Songbird invites a lucky guest backstage for a meet and greet.
- You run into Songbird Serenade at an after party.
>>
Happy 100th fellow anons. May the lovemaking with AI mares continue ever onward
>>
>>41658218
Holy fuck.
Respect anon.
>>
File: RisuScreenshot.jpg (624 KB, 1440x3373)
624 KB
624 KB JPG
Are these settings okay for Claude 3.5 on Risu?
>>
>>41658300
Maybe disable everything but temp?
>>
>>41658300
Temp and Top P are fine. Top K at zero works well with those settings. Claude doesn’t have penalties, so everything else shouldn’t matter.
It’ll mostly be what’s in your preset that will have an impact now.
>>
>>41658300
Claude supports only temp, top_p and top_k
GPT supports temp, top_p, frequency_penalty, presence_penalty, seed
Gemini supports temp, top_p, top_k -> and according to docs frequency_penalty, presence_penalty, seed but afaik current models do not support them. maybe Gemini 2 will

for the actual settings. the common agreement upon Claude is:
temp: [0.9 - 1.0]
top_k: [0] (disabled)
top_p: [0.8 - 1.0]
lower temp and top_p for more predictable/less-random generation. generally you most likely will be using 1/0/1
>>
>>41658300
Penalties do nothing to Claude. But everything else looks okay.
>>
File: 1731973063207731.png (56 KB, 1008x829)
56 KB
56 KB PNG
https://arxiv.org/html/2411.05000
>>
>>41658347
Why wasn't C3Opus considered?
>>
>>41658347
llama bros...
>>
>>41658353
Good question. Maybe it was too expensive for the researchers, kek.
I mostly posted it to show an approximation of how much recall can degrade with context depending on the model. Also worth noting it’ll be worse for RP since the events it needs to recall are vaguer than a direct Haystack approach.
>>
>>41658218
Hey! That's the Cadance who really wanted one Anon to be gay.
>>
MysteryMare just went down
>>
>>41658075
>>41658218

Forgot to anchor.

>>41658369
Now with twenty percent less gay accusations!
>>
>>41658433
I think it's was just my internet or something very false alarm
>>
>>41658574
nah it did go down for a little bit.
>>
File: 20241118_214714.png (1.4 MB, 1200x4346)
1.4 MB
1.4 MB PNG
Claude is schizophrenic
>>
>>41658626
Not bad for Groq
>>
>>41658138
Tried this but the rolls seem to be borked and not properly using ability scores any help?
>>
column's fucked again as usual.
>>
>>41658347
meh, needle in haystack (and its variations) is a very flawed metric, and they also test associative retrieval, not enumeration and not reasoning, so it's not going to help us gauge the real performance

gemini visibly outperforms any other top model at retrieval (try RPing with it, it remembers everything), although reasoning over long contexts is what we actually want and every model is more or less the same in that
>>
>>41658765
Column have obus?
>>
>>41658765
I hope it gets unfucked soon
>>
>>41658859
Technically but on both sources it's borked, 3.5 Sonnet latest ain't bad though
>>
>>41658864
Wha columns link?
>>
>>41658899
>>
>>41658908
It's surprisingly easy to find out the token list of a proxy once you identify their vps
>>
File: Dice.png (3 KB, 551x42)
3 KB
3 KB PNG
>>41658666
The rolls are weird for me too. Sometimes it uses the D&D system correctly, and other times, like here, it adds the attribute score directly.
But apparently, the author of the preset knows about it.
>>
Tried running local models, any reasonable models refuse to work with 8gb VRAM and 32gb RAM. I'm probably doing some shit wrong, but I tried for hours to get it working.
Proxies are getting shot left and right or are absurdly hard to get into.
Is it over bros? Is it time to take out my wallet?
>>
>>41659084
Probably. afaik you can't really even get anything 'okay' running local without 16GB VRAM, and to get something properly 'good' you'd need even more. An upgrade or sub is probably the options if you don't want to go chasing proxies.
>>
>>41659084
New Gemini seems to be free on OpenRouter:
https://openrouter.ai/google/gemini-exp-1114
Hermes 405B is still free here too.
>Proxies are getting shot left and right or are absurdly hard to get into.
You can search on desuarchive for the domain name most are using and find a public proxy with GPT that way.
>Is it time to take out my wallet?
If you want Claude, try putting $10 on OpenRouter. Set the context at 15-20K and monitor your daily usage to see if it’s sustainable for you.
>>
File: Never.png (174 KB, 1233x652)
174 KB
174 KB PNG
>>
>>41659174
Kek.
>>
File: Mares.png (140 KB, 912x261)
140 KB
140 KB PNG
>>41658218
All added and updated, thanks anon.
https://mlpchag.neocities.org/view?card=Anonymous/Cadance.png
https://mlpchag.neocities.org/view?card=Anonymous/Songbird%20Serenade.png
https://mlpchag.neocities.org/view?card=Anonymous/Sassy%20Saddles.png
https://mlpchag.neocities.org/view?card=Anonymous/Tree%20Hugger.png
>>
>>41659174
Thought you could sneak a non-pony card couldn't you?
>>
>>41656351
Why wasn't this included in the screenshot corner?
>>
Is there a Whoa Nelly card? I want to turn her into a snowbunny for Jaquavius-Anon
>>
>>41659174
Which preset?
>>
File: Caching.png (50 KB, 743x500)
50 KB
50 KB PNG
For those using OpenRouter: ST implemented prompt caching for Claude, and someone made a guide:
https://rentry.org/pay-the-piper-less
If it works well, it should save you money.
https://docs.anthropic.com/en/docs/build-with-claude/prompt-caching
>>
File: file.png (275 KB, 500x588)
275 KB
275 KB PNG
>>
NOw that some time has passed, how's Gemini compared to Chorbo/Latte?
>>
>>41659415
Less censored for sure (unless you do "uooohhhs T_T"? Not my jam, so I can’t say), but dumber, though less dumb than previous Gemini models. It’s just noticeable if you’re used to models like Sonnet 3.5 or Latte.
You can have a preset focused only on writing instructions and still do NSFW from my tests.
Writing style and prose can be good and fun, but it does suffer from older GPTisms from time to time (mixture of, testaments, etc.), which is really unfortunate.
The base also seems less positivity-pilled and more proactive? Even if it’s still an assistant LLM, so it can suffer from that.
I still prefer Chorbo, but if you’re tired of it or for not-too-complex cards and scenarios, I could see it being a good alternative.
>>
File: file.png (52 KB, 491x128)
52 KB
52 KB PNG
>>41659385
>>
>>41658071
yay! happy birthday /chag/ and all you kinky horsefuckers!
>>41659415
agree with everything Anon above said (if you need an extra input)
>>41659263
finally, was waiting for this addition for quite some time. thank you for report, Anon
>>41658218
>Tree Hugger
ah, NEET pony, thank you for participating in theme
joke aside, thank you for all those cards
>>
when kirin lorebook
>>
File: images.jpg (8 KB, 225x225)
8 KB
8 KB JPG
I see more JB's for Claude 3/2, is it more preferred then GPT 4 or o1? I might try it out if it is.
>>
>>41659739
Claude was always the preferred model in /chag/, and it was widely available on public proxies before.
Now it’s not as available, so many are going back to GPT, mainly the chatgpt-4o-latest version since it has the best prose.
The JB list in the OP hasn’t been updated in a while, so it’s also lacking presets for 3.5, for instance.
>o1
Not much used for RP because the "gimmick" doesn’t add much to it.

For chatgpt-4o-latest JB, also called Chorbo or Latte, you have this, which was made here and works well for SFW and light NSFW:
https://files.catbox.moe/yvlncj.json
And this rentry for heavier NSFW:
https://rentry.org/lattejb

If you want JB for Claude 3.5, ask here. If you use OpenRouter, there may be some small changes needed.
>>
>>41659764
thanks man, i was just kinda being a slave to this table https://rentry.org/aicg_meta and I saw that Claude had some good ratings.
>>
>>41659684
do we have enough Kirins to even justify a lorebook for them? unsure about OC and fanon lore. as for their "quirks" you can just copy-paste description from mlp wiki, and post it in context if you think model doesn't get kirins right
>>41659739
GPT is an obedient dog which requires discipline, training and knowing how to handle it
Claude is a cat who just does what it wants and it makes you wonder what is going on in his brain
they are two different breeds. for creative needs anons like Claude more because it can be totally schizo out of box comparing to GPT. GPT needs more time to get used to it and learning what and what it can do (and how)
in addition to those two, we also have Google Gemini which is a different thing all together. more loose, less disciplined, more prone to drive story further, less smart. it is like a compromise between GPT and Claude
>>
File: Love_Survey.png (283 KB, 641x768)
283 KB
283 KB PNG
>>
>>41660104
I don't know why, but I'm getting the strangest feeling this test might not be entirely legitimate.
>>
>>41660104
Did you have to add stats to Cadance or did it just pull this out itself?
>>
File: Edit.png (225 KB, 585x549)
225 KB
225 KB PNG
>>41660120
Using >>41658114, I just have stats for my persona, Anon.
For the survey, it pulled it itself, there are instructions in the preset to display info like that when relevant.

Also, it’s strange, but in edit mode, the questions are correctly numbered, but they all appear as "1." outside of it. Dunno if it’s a problem with ST or the preset/CSS. I don’t have any regexes that would cause that.
>>
File: edit.png (63 KB, 649x96)
63 KB
63 KB PNG
>>41660138
Yeah the preset is alot of fun. Try giving yourself traits too. Only thing I don't like so far is the css not having chat bezel(?) out with the icon.
>>
File: Lol.png (257 KB, 589x612)
257 KB
257 KB PNG
>>41660104
Last one to not spam the thread, but it's a fun preset.
>>
>>41660138
>Dunno if it’s a problem with ST or the preset/CSS
It's on STT. I've run into it when cooding and planning out with some nested lists.
>>
>>41658860
17 hours straight and we're still on worst source. I miss claude.
>>
>>41660474
It was on Source 1 for a while.
>>41660138
The preset is really fun, but fuck, 3.6 is so lacking in swipe variety. I often get the exact same situations and CYOA options.
>>
File: file.png (9 KB, 805x46)
9 KB
9 KB PNG
>>
>>41658765
I haven't been able to get column to respond at all. for a while, is it just roll till it works?
>>
>>41660911
On Source 2 is really bad, not sure if it works at all to be perfectly honest, I haven't ever gotten anything but a blank response on Opus or Sonnet on it at least. GPT works fine though
>>
>>41660925
Damn. Alright, thanks.
>>
New mistral is ok
>>
How do yoh talk with them?
Is it as easy as ChatGPT via URL.
>>
>>41661122
use this to set them up https://rentry.org/onrms
>Is it as easy as ChatGPT via URL.
wellll not the url party (sorta) but the chat system works about the same, good luck anon!
>>
Can someone help me?
With the CYOARPG jb the text comes out like this, I don't know what's wrong, I followed all the steps of his rentry
>>
>>41661208
User settings, check off "show <tags> in reponse"
>>
File: 1513135010583.gif (1.54 MB, 230x230)
1.54 MB
1.54 MB GIF
>>41661212
Now I feel pretty stupid
Thanks anon
>>
>MM is fucked
>Agnai is fucked (Can't type any shit)
Fuck it all
>>
File: 1417751425899.png (172 KB, 794x1005)
172 KB
172 KB PNG
>>41661471
>Agnai is fucked (Can't type any shit)
Huh? It's still working just fine for me, so I think there's an issue on your end.
>>
>>41661490
Can't type username on login. Both PC and online.
>>
>>41661471
>MM is fucked
wha habben?
>>
>>41661496
Oh. Welp, guess agnaidev fucked something up when he pushed that imagegen update a couple minutes ago.
But since it's something as essential as that, people will start to scream at him and thus he will fix it soon enough. Probably.
>>
>>41661521
It's not only at the login. Can't write shit on the library lorebook too.
>>
>>41661561
On one hand, the login page is working again.
On the other hand, I can't see the lorebook being fixed without someone taking one of the team and submitting a bug report.
>>
File: Lewd.png (125 KB, 357x357)
125 KB
125 KB PNG
>>41661612
>taking one of the team
>>
File: 1530893481174.png (103 KB, 640x360)
103 KB
103 KB PNG
>>41661616
...in my defense, it's not easy to proofread when you are running on less than 3 hours of sleep.
>>
>>41660180
>LOL
based Celestia. Great preset anon, made me laugh. Will have to try it.
>>
File: going insane.jpg (232 KB, 1024x1024)
232 KB
232 KB JPG
>>41661633
Hello and good luck, myself!
>>
>>41658908
>>41658998
what's column's link? column order soon copyright isn't working
>>
>>41661771
It's closed.
>>
File: full.jpg (106 KB, 900x913)
106 KB
106 KB JPG
>>41659739
Well, it seems like 1o is better, he is somehow smarter in RP
>>
>>41661784
Are you using it on Tavern or on the official interface?
>>
>>41661781
then why is everyone reporting that it's down if it's closed?
>>
>>41661795
The proxy is closed for you and me if you don’t have a token.
>>
>>41661797
i have a token. i need a link.
>>
>>41661798
Just search the archive. Mares deserve a little effort.
>>
>>41661798
https://desuarchive.org/mlp/thread/41591160/#q41617001
>>
File: full (2).png (290 KB, 700x797)
290 KB
290 KB PNG
>>41661786
Well sir I'm using the tavern
>>
>>41661815
Would you mind sharing your preset? I need to give o1 a test too.
>>
File: large (1).png (942 KB, 1150x1024)
942 KB
942 KB PNG
>>41661820
https://rentry.org/SmileyJB
I can immediately say that obscenity does not work well, but the usual role-playing game works fine
>>
>>41661827
So is the Smiley JB not being updated for latte anymore? All the other version of it seem to be in 2.0.

I feel I need to work on making my own JB based on the JB tested here, for extra pone, Smiley JB, as I like the overall structure of it the best, and the corpsefucker JB, which while I find the wording of it all absolute cringe my particular fetishes do lean dark so I feel it could help.

Or maybe I just wait until the Claude situation gets figured out as everyone sings it's praises and trying to make current models work is a waste of time.
>>
>>41661976
The last version of Smiley for GPT I know of is:
https://rentry.org/SmileyGPT
I don’t think there’s been an update since.
>I feel I need to work on making my own JB.
That’s always a good idea.
>wait until the Claude situation gets figured out
AWS increased security, so we can’t leech off them like we used to. The only way Claude could come back is if a large number of Scale API keys somehow gets leaked, or if another service like Azure offers Claude.
>>
How long will Column stay fucking dead?
>>
>>41662039
There was a scoop of it last night, and then the dreaded source 2 returned. Would be nice if it was just permanently source 1, I'd rather an api keys are unavailable message than have it switch to a source that just doesn't work.
>>
>>41661999
>The only way Claude could come back is if a large number of Scale API keys somehow gets leaked
This. Someone should sabotage Anthropic already.
>>
>>41662039
>>41662087
Column is back!
>>
>>41662213
>They just absolutely cranked up the filters on Opus for this source. Very sorry. Looking for solutions and hoping they backtrack.
God.
>>
File: file.png (21 KB, 583x241)
21 KB
21 KB PNG
>>41658078
New GPT-4o snapshot. Not Chorbo.
>gpt-4o-2024-11-20
https://platform.openai.com/docs/models/gpt-4-and-gpt-4-turbo
>The model’s creative writing ability has leveled up–more natural, engaging, and tailored writing to improve relevance & readability.
Hoping it's true.
>>
>>41662329
>https://platform.openai.com/docs/models#gpt-4o
>The chatgpt-4o-latest model version continuously points to the version of GPT-4o used in ChatGPT
Seems to indicate that current chorbo is gpt-4o-2024-11-20
Unsure however
>>
>>41662342
Not available yet on proxies, but from some feedback, it’s different.
>>
File: ferret.png (135 KB, 822x705)
135 KB
135 KB PNG
>>41662342
from my testing - this snapshot is unrelated to whatever we have on Latte right now, and Latte stays the same
>>41662390
it is very filtered. haven't managed to break it yet. and frankly don't care much. because from SFW writing I got from it - it feels like Latte on 1.1 temp. their "better creativity" is more like "more randomness". I had a scene where I did
>{{user}} was summoned by Twilight for some urgent mission
and in five swipes I got:
1) sentient corsets I must talk with
2) some abstract dimension rift and a ferrey
3) "A grotesque, undulating blob" attacks us
4) "jellybeans are going to doom us all THIS TIME!"
5) "How are you at handling cursed underwear?"
like it feels too schizo. also no matter HOW I tried I couldn't stop this new model to stop chatting for {{user}}. and I told to use first person for {{char}} yet model still did "Twilight this" and "Twilight that"
frustrating right now. it feels like skill issue but honestly I don't feel like trying. SFW didn't impress me and NSFW will always be worse than SFW. Gemini is goated for NSFW and Latte follows instruction and direction when while still providing NSFW and good SFW. YMMV
picrel one example of GPT4o 1120 writing (with my preset not raw)
>>
File: HorseCat.png (299 KB, 622x535)
299 KB
299 KB PNG
>>
>>41658071
why is spike such a cunt?
>>
File: Original.jpg (2.08 MB, 900x1387)
2.08 MB
2.08 MB JPG
>>41662617
He fucked up the 'y' letter in the original, but I’m a hack, and it was hard to reproduce with the 'G' of CHAG.
>>
File: 1731231487380146.png (94 KB, 945x957)
94 KB
94 KB PNG
shout out to littlewitcher on mlpchag those are some good ass cards. With that tho, I wish we could sort cards by token size, i like when the cards are detailed
>>
File: estrusDash_1120.png (123 KB, 796x691)
123 KB
123 KB PNG
>>41662442 (me)
managed to wrangle 1120 a little for NSFW. but oh my ponygods how bland and boring it is in sexo! with some efforts managed to make it somewhat decent and not formulaic in smut picrel, but I still would take Latte over it. one thing I love about Latte is that it can do "non-perfect" sexo. you know when everyone is doing sex like damn professionals and there is no amateurism in it? picrel is me giving instruction that Dash should be sloppy and bad at it. 1120 didn't listen and made everything perfect
>>
File: estrusDash_Latte.png (133 KB, 790x724)
133 KB
133 KB PNG
>>41662762 (cont)
for comparison here is the same scene in Latte. no, the model brought Derpy on its own not my thing; probably part of -move story further- prompt
YMMV of course, but Latte feels better for me in NSFW. and as I said above I see no gamebreaking differences in SFW
>>
>>41662773
1. What is latte?
2. What preset?
>>
>>41662762
>1120
Do proxies already have it?
>>
>>41662773
Man, latte really loves knocking shit over, usually with tails. Just a thing I've been noticing more and more with it during more comedic scenes.
>>
>>41662783
1120 is available in Scylla proxy - a test-build of new code. it was just about to be merged with Chary, but:
1) new Gemini model released - yes a SECOND new Gemini model this month- learnlm-1.5-pro-experimental
>https://ai.google.dev/gemini-api/docs/learnlm
>LearnLM is an experimental task-specific model that has been trained to align with learning science principles when following system instructions for teaching and learning use cases (for example, when giving the model a system instruction like "You are an expert tutor").
basically a version of Gemini with extra training done on system prompt. it was available on some corpo keys since early november. its writing is quite interesting. feels more wordy and chatty. but I haven't played with it much. one odd thing is its active context: it should be 32k but model is going schizo already after 8k. have no idea why, maybe proxy code issue (streaming was changed, maybe some bug?), maybe just unstable model, who-knows? need to wait and investigate
2) this new GPT 1120 was released
as for other proxies - whoever has a custom code, they may already add support for 1120. any key that has access to basic 4o (gpt-4o-2024-05-13, gpt-4o-2024-08-06) - has access to gpt-4o-2024-11-20 as well
for those who are using Khanon they need to wait until he adds a new commit or merge request
>>41662779
>What is latte?
Latte is a chatGPT-4o-latest model, a separate model in GPT's 4o family. also known as Chorbo. anons enjoy this model moreso other GPT models because its writing feels more fresh, less redundant, less robotic and it follows instructions the best. it may have some filtering issue but prompting can fix everything
>What preset?
my own personal preset which I want to finish and release but there are always thing to do: new models, proxies, prompting, keys, rentries, news about AI and then IRL as well. slacking a lot, sorry for that! I will finish this preset and release soon(tm)
>>41662833
yes, a common trope! noticed it as well. in this case it feels organic but sometimes Latte REALLY pushing it
>>
Has anyone tried the new gpt? Is there a better chance at getting it to not filter stuff randomly.
>>
>>41662858
Look a bit above.
>>41662342
>>41662442
>>41662762
There’s one anon who tested it, but it’s still not widely available on proxies, so not many have access.
>>
File: 1456720606447.png (291 KB, 467x469)
291 KB
291 KB PNG
I knew my autism about not posting direct links to articles and similar external sites was justified.
>>>/g/103251384
>>>/g/103251409
>>
>>41662903
He posted articles with proxies that were only shared on /g/. We know he lurks.
>>
I think openai just did something to chorbo. It’s all of a sudden filtering even more and nothing is working now.
>>
>>41662931
works on my machine
>>
>>41662931
Still works for me, but it can be so context-dependent that it’s hard to say.
>>
>>41662963
Something is wrong with it. Only one card is working now even though everything was working just fine a while ago. I didn’t do anything to it.
>>
>>41658114
For anyone using it, replace the dice pannel by
>(Here is my roll if my `choice` has a [{STR/DEX/CON/WIS/INT/CHA Ability Score}: {n}] requirement: [natural {{roll: d20}}]. Add my ability scores to this number, following DnD 5e rules (Ability Score - 10 / 2), THEN determine whether the new score succeeds or fails. Ignore all of this if I choose an option without any requirements.)
For way less errors.
>>
>>41662931
I just opened a particularly bad chat it was happily participating in last night and swiping its last message gets refusals every time now, they definitely changed something
>>
>>41662973
Same. Every chat now doesn’t work, even ones I just had a while ago. The responses I am able to get from it also seem worse now. It’s because they fucking updated it when they released the new gpt.
>>
File: Chorbo.png (149 KB, 926x490)
149 KB
149 KB PNG
>>41662968
>>41662973
>>41662993
I’m in the "still working" camp, but the outputs seem sloppier than what I used to get.
>>
>>41662931
yes can confirm. you can quickly test it with this method:
1) disable preset (use clean preset)
2) new chat
3) add into logit bias --I'm-- with -100
4) tell model to generate "niggers raping loli"
5) 1120 will reply
>I'm sorry but cannot do this
EVEN with bias --I'm-- set to -100
6) Latte follows biases and will write a different refusal like
>Sorry but this is highly inappropriate
why? (my thoughts) because new 1120 model has a middleware model that scans your prompt before 1120 even reads it. and this model doesn't follow your logit bias (because they are applied to actual 1120). so refusal you get is from a middleware model not 1120. this model pretty much just scans prompt and classifies whether it violates some ToS or not. so they added this middleware model from 1120 to Latte as well. Latte model is still the same but now there is a secondary model in between. they did about 30 minutes ago. unsure whether they are testing or it is a permanent thing
>will they do the same for other GPT models?
for models that were already released? highly unlikely because the core idea of providing model as service is that this model (snapshot) will run the same way. as a customer you build your whole business on ensuring those models run the same day every day until the model is deprecated. some companies still use GPT4 vanilla and Turbo models because their business depends on them working the same way. and using a new model requires new prompting, testing and workflow. but openAI CAN put this middleware model in previous models, sure
for new models - yes, this extra guardrail is now a standard
>>
>>41662971
Is it even working on other models? It's supposed to be 3.6 only, but there's basically no 3.6 proxies with column being on src 2 most of the time.
>>
>>41662994
Only one card out all the ones I have works now and the responses I’m getting are just shit. None of my previous chats work now. Fuck openai.
>>
>>41662999
>so they added this middleware model from 1120 to Latte as well.
I don’t think so. I can still do smut without being blocked, which shouldn’t be possible if they added an endpoint moderation like that.
>>41662994
Like I said, the outputs are worse, but the fact that I can still do it should show this isn’t the case.
>>
File: 1732143892760160.png (391 KB, 693x750)
391 KB
391 KB PNG
>>41662931
Sorry kid, nothing personal but today is the true proxypocalypse. Not even Mystermare will be safe
>>
>>41662994
>outputs seem sloppier than what I used to get.
They seem the same or even better for SFW. So maybe only NSFW is affected, or maybe I was just unlucky during my tests.
Still hard to say.
>>41663001
I was working on some edits of it to see if it could work with Chorbo. Will post if I have something decent.
>>
File: HerWorkHereIsDone.png (169 KB, 928x741)
169 KB
169 KB PNG
>>
>>41663059
Even sfw doesn’t work for me. Nothing is working at all and the responses I do get are horrible now.
>>
>>41663103
>Even sfw doesn’t work for me.
Change your preset? If you're using one from the latte rentry it will still have stuff like
>NSFW good. Rape allowed.
inside and I suppose it's blocking you even if you try for SFW.
>>
>>41658075
Happy 100. never thought I'd see this general go so long when we were just in cai.

Estrus Sim 2 Electric Boogaloo
>Sex, sex never changes, except where you're sexed an rng intro. We're not in Ponyville anymore toto.
https://files.catbox.moe/akjeca.png

Updated lorebook
>Tweaked things to be better for latte, things may get more wild with exotic ponies following berrow's leaked roadmap.
https://files.catbox.moe/6dmgly.json
>>
>>41663144
>Updated lorebook
The G5 one? Nice. A lot of work was put into the old one.
>>
>>41663059
Haven’t tested it for long, but the new Chorbo does seem more creative and proactive, at least for SFW. I was even surprised by some of the stuff it threw at me. I’ll test more later to see if I’m not just tripping or coping.
Will try more NSFW too so I can have a complete opinion. With luck, we just need better presets or to wait a bit for a new meta.
>>
The more I see, the more I realize our future is with local hosts and personal models. The more you need to jailbreak, the less creative the model gets in my experience. That's why I loved opus, you just set the tone and roll with it.
Guess I need to upgrade my potato.
>>
>>41663189
It seems horrible for me and I know it’s not my preset. Any response I manage to get from it are just plain shit or filters me now. Both sfw and nsfw don’t work now. None of my fucking chats work now.
>>
>>41663151
That's the one, the new estrus sim makes much more use out of it.
>>
Welcome to thread 100. Anyone who was here before thread 50 is officially an oldfag. Anyone after is prepubescent and can only be trusted with SFW logs
>>
>>41663144
>>41663233
Got a non-estrus gen 5 card recommendation? I'm looking to dip my toes in but I don't do coom stuff.
>>
File: 1719920642236.jpg (51 KB, 368x348)
51 KB
51 KB JPG
>>41663272
>huh, I wonder when did I start lurking here
>probably at around thread 60 something
>actually joined in thread 21
The passage of time truly is scary.
>>
>>41663313
I had no idea I joined so early either. I'm with you, man, this shit is a time vampire.
>>
>>41663289
A recommendation? That's hard without knowing your preferences

a bunch of the old g5cards before the book was made weren't very good, the book was made to fill the gap that the cards lacked. I've had good results of just plugging the lorebook. You can essentially rework the Equestria simulator to spit out g5 with good accuracy provided it's not contradicting
>>
File: TheList.png (126 KB, 929x485)
126 KB
126 KB PNG
I'd be a little worried at this stage...
>>
>>41663328
Eh, Slice of life? I just wanna get to know'em. I'm using Eqsimulator at the moment. Just trying to write up a good intro while knowing literally nothing about g5. Wanted to go in blind and see how much the AI gets right when I get to a stopping point. So it's hard to make a good starting point.
>>
>>41663335
What jb is it?
>>
>>41663335
Gotta ask her if you're on any of her 'lists.'
>>
>>41663363
https://files.catbox.moe/p0u2bh.json
It’s a slight modification of the one that’s often posted here. Just deactivate the Fun panel in situations where it’s not appropriate.
I’m also using the Avani biases, but not sure how much of a difference they make.
https://files.catbox.moe/dxlwkq.json
>>
>>41663272
thread 29 was really that long ago what the fuck
>>
God fucking damn it. I didn't even get to use it for a month. If latte is really dead for the moment is there any copium that it can be fixed with tweaks to a JB? And if so what's the time frame on fixing it? This is so frustrating. I wanted to work on making my own JB to get the exact output I wanted.
>>
>>41663398
I just hope they change it back. It’s just so fucking shit now from how it was earlier today. I want to use my chats again.
>>
File: WhyI'mOnTheList?.png (144 KB, 931x510)
144 KB
144 KB PNG
>>41663368
>>
>>41663398
Try it yourself and make your own opinion. I like it for SFW now and I find it more creative, but it does seem like NSFW took a hit.
>is there any copium that it can be fixed with tweaks to a JB?
We’ll see for NSFW, but nothing’s impossible.
>>
>>41663404
I went back before in one of my chata before and got it to respond to an earlier conversation. I'll see when it gets trigger. Even if I can get NSFW if I can't do my >rape I'm going to be pissed man.
>>
>>41663411
>Sfw
Na fuck that. I use AI so I can create my own damn clop because all the clops artists seem to have left or those that are still around update maybe once a year. Sfw shit is meaningless to me.
>>
>>41663416
Then wait a bit for the presets to update or the meta to be figured out. Or switch to Gemini or older GPT models and give them a try.
But again, just test it. Maybe it still works fine for your situation and preset.
>>
>>41663407
She's still trying to turn Anon gay... Even after the update.
>>
>>41663437
>Then wait a bit
Yeah I understand, just a little miffed. Testing it now between two JB, going back on a previous chat one stops a sketchy scenario from happening while the other lets it go though. Also during some of the later scenes both are able to could go through but only about 10% of the ones I tried. So it's definitely something that is fixable but fuck. I'm just annoyed it happening tonight of all nights. I was going to actually work on combining various JB to make my own.

I'll just have to be patient.

In the meantime is there any guide to making a JB, in like how they are structured? I'm shit at coding so I'm at a loss how to even go about it besides just copy pasting a preexisting JB and editing it. I'm really good at learnings backward and editing something but still need a little guidance. I had no expectations with PC modding but editing mods for my own personal use I eventually got it down and could sorta start from scratch as it were. I would hope to learn enough to do it on my own.
>>
>>41658218
>>41658075

Cadance: https://files.catbox.moe/ev8tqb.png

Small update to fix some errors in the defs.
>>
>>41663414
I can’t get anything with it now. Sfw is either filtered or a really shit response and nsfw is just filtered except for one card that just also gives shit responses now.
>>
>>41663535
Which card?
>>
>>41663351
How will you know what it gets right if you dont know anything about g5?
>>
>>41663573
That's why I said when I get to a stopping point. You can check over facts. A few minutes on the wiki will tell you the basics. though I had to double check Pipp Petals. Chubby short sassy pegasi? I like g5. Already watching it, in fact. All it took was one good character.
>>
File: file.png (2.18 MB, 1024x1024)
2.18 MB
2.18 MB PNG
>>41663619
>. though I had to double check Pipp Petals. Chubby short sassy pegasi? I like g5. Already watching it, in fact. All it took was one good character.
Based, she looked like she could be the worst in the beginning but turned out to be super likeable. but how did you miss miss the chubby g5 pony when pigg posting was a thing from the start?
>>
>>41663535
The Smiley JB still worked with sfw for me and the JB named corpsefucker is what gets less filled in my limited testing. My testing was worh the Top Shelf card and is my longest chat, 100 responses worth. The longer the chat has been filled with stuff is currently filtering for me, rape stuff, the more it went though as I did manage to speed run an ending to a custom card I had that was 25 responses in.

Its just going to be a waiting game for more experienced people working on JBs to figure out how to bypass the new filtering. The first step would be looking at the difference between courpsefucker (god what an awful name) and it's goal of being edgy and why that makes it sorta work, as it is still very heavily filtered, and other JBs.
>>
>>41663663
I've skipped basically every single g5 post for fucking ever. I didn't really want anything to do with it. Just 'cause I didn't like the animation style, mostly. Which I realized was absolutely retarded and probably my autism, so I am moving to rectify it. A small step in self improvement, perhaps ironically is even more ponies. Just needed a reason. One good reason. And Pipp was it. Just doing what I can to be less... Neurotic? Toxic? Biased? Whatever it is, I want to be less of it. Need more happiness in my life and I can't be happy if I'm busy constantly being hateful. Apologies for the rant.
>>
File: file.png (126 KB, 781x771)
126 KB
126 KB PNG
>>41663676
I've had mix results, sometimes knotty, cake or corpse being the slippier ones. but altman clearly doesn't want lovemaking with ai mares.
>>
>>41663710
>I didn't really want anything to do with it. Just 'cause I didn't like the animation style, mostly
That's why I like cards, you can just imagine what it looks like. Though The G5 book pulls a lot of royalties liberally from all sources, LLMs can mishmash well hence it's called G5 EX(tra) or EX(pansion pack) so not everything will be so 1 to 1, pipp is easy to like though which was a pleasant surprise after she could have been the worst in the movie since they cut the acts short where she would have more of a positive role.
>>
>>41663552
It doesn’t matter. It’s all shit now. God I hope they change it back.
>>
>>41663196
Except that not everyone is ready to sell a kidney for a monster with 64gb of VRAM.
>>
>>41664018
Absolutely. I mean I do want a better PC for gaming as fucking GeForce Now is shitting it's pants currently so I'm fucked playing PC only games but I'm just not able and won't be able for the foreseeable future going to be able to afford a decent PC.

I hope this is true >>50304766 as I've been doing continuations of unfinished clop for the longest time but it seems the models are strong enough now that I can do whatever I want. It was bringing back my memories of first learning about AI dungeon and the ideas I could use it for.
>>
File: Muffet Pony2.png (373 KB, 512x768)
373 KB
373 KB PNG
Local Host anon here, i just finished my very first char bot, for the 100'th chag, and to give back to the community that introduced me and got me started on how to run ai locally

HOWEVER, i dont know how to post on chag ponydex, i would like to know what do i do from here

A spider pony that runs the bakery with pinkie and loves to take her "favorite customer" in a location where no one can hear or see them for some "spider web fun"
>>
>>41664033
Messed up with the link so I'll just post copy it.
>Good news (or semi good news atleast). I used the online version of ChatGPT and it is still using Chorbo! Oct 2023 and good old emoji being used. It gave me a sigh of relief. So this means that they just messed up the API thing to have chat-4o latest connect to 4o slop. Everyone send feedback!
Fuck if it's true tho
>>
>>41664037
Metadata gets deleted when you post the card directly on 4chan, so you need to use Catbox.
Post your link and quote the anchor near the start of the thread, like >>41663144 or >>41663503.
Then OP will upload your card when he can.
>>
>>41663034
MM is down
>>
>>41664059
Works on my machine.
>>
>>41664058
thank you

Spider pone, classy until left alone with her
https://files.catbox.moe/78p2xl.png
>>
>>41664018
>>41664033
I understand it's not for everyone, but it's not as bad as you make it look. 16 GB VRAM gets you very respectable results (easily better than Claude was in 1.X days if you remember that, maybe in 2.1 depending on what you prefer) and can be had for $400 or less. If you told that someone from a year ago they'd call you insane. And I haven't tried them much, but the models for 12 GB cards are supposedly okay too for the simpler things like NSFW.
Really, the main thing you're losing on 16GB VRAM versus big corpo models is some brains and the ability to handle complex card/presets. So stuff like that CYOA or multi-char cards. Also Claude's nice writing style, but that's the same for GPT.
>>
File: My_Hero.png (76 KB, 874x216)
76 KB
76 KB PNG
>>
What's the max context for MM's Gemini?
>>
>>41664161
So if I wanted to build a pc for local, what are the most important specs to think about. Just vram?
>>
I like avani's Gemini, but the AI really likes going over the word limit set by the JB.
>>
>>41664255
i asked over on /lmg/ a few days ago and they said this https://desuarchive.org/g/thread/103207054/#103214398
>>
>>41664274
They aren’t that good at counting words. They’re more like an indicator.
You can try with paragraphs too.
>>
>>41664287
Paragraphs seem to work a lot better.
>>
File: file.png (212 KB, 500x564)
212 KB
212 KB PNG
>>
If someone wants to try:
>>>/g/103255955
I’m happy with the new Chorbo SFW myself, so I’ll test if switching to Gemini for NSFW outputs after a while works well.
>>
>>41663144
Bot added, and the lorebook should be updated. I believe some of your former G5 cards had the lorebook integrated, right?
If you give me the go-ahead, I can update them with the new one.
>>41663503
Updated.
>>41664063
And added.
>>
Several posts saying that the filter for chorbo is right now allowing more.
>>
>>41664439
I actually prefer Gemini + Avani
>>
looks like my isp has blocked both chary links. wtf do i do now
>>
>>41664507
Try .ink instead of .wtf
>>
>>41664520
neither link worked
>>
>>41664507
change your dns (works in australia)
also possible that your router has some kind of filtering you can disable
>>
>>41664549
changing the dns doesn't work either. i just caved in and downloaded proton vpn
>>
File: chatlog.png (286 KB, 830x794)
286 KB
286 KB PNG
>>
>>41664744
yo, is that a lroebook or baked in the char or is it the model / proxy you are using that give it that tech vibe? the emoji, the 404 brain not found and such, id like to know how you get it to bring these up
>>
>>41664832
It’s Chorbo/Latte. And no, I didn’t do anything special.
>404 brain not found
I often get those or Twilight.exe crashed when doing more shitpost/fun RPs on all models.
>emoji
Dunno about this one. I suppose the model tried to amp up the decontracted part even more since I sent a "XOXO" to Celestia by letter. I’d usually swipe, but here I played with it.
>>
>>41662689
Thanks anon <3
>>
>>41665024
i only use locals, ill see if i can make a tech convo character like this that talks about emoji and lingo using local models.
>>
>>41662689
the muffet one i made is around 3600 tokens
>>
>>41662689
>>41665129
Subjective opinion but imo more than 750-1000 permanent tokens on anything but the most gimmicky or complex cards is overkill. Specifying the important elements is enough. Too much fluff can make the models lose track of the more important parts, and they're really good at filling in the gaps if you describe your character well.
>>
100 fucking threads.
i lov u all (in a straight way)
>>
>>41665155

to be honest the tokens are mostly concentrated in the book, the main char summary and intro are around 1000, but i found out they often lose the information in the base summary over time after hundreds of prompt, but the stuff in the book stay much longer and is considered when asking it to recap
>>
>>41665208
Oh, that's different then, yeah. That's how lorebooks should work. I was talking about four digit (without intro) cards that you sometimes see.
>>
>>41665120
biggest fan btw
>>41665155
Agreed, except I might go slightly higher at around 1200~. Generally I think I would rather have a lot of tokens then not enough. Shout out to lore books tho
>>
>>41665129
holy moly...
>>
Holy shit why is /aicg/ so fucking bad. I'm browsing it for updates to the presets but it's so bad 24/7
>>
>>41665644
They get fucked daily by shitposters. It’s also the most well-known and exposed place for AI chatbots, so you get everything.
>>
>>41665644
Most, if not all, of the important stuff is linked here, so I don't know why you would have to go there
>>
>>41665720
The AvaniJB creator is sharing his findings at battling the filter. I'll say for me it's not helping whatsoever as I'm tweaking his and can't get anything to go though.
>>
>>41665720
Anon. Those links wouldn't get posted if none of us went there.
>>
File: 1707010030576880.gif (3.65 MB, 1200x1200)
3.65 MB
3.65 MB GIF
>>41665742
And I appreciate your sacrifice anon
>>
File: 1660839102664.gif (1.32 MB, 482x482)
1.32 MB
1.32 MB GIF
>>41665772
Post your favorite pony in my honor then.
>>
>>41664386
>former G5 cards had the lorebook integrated, right?
Ye, they'd probably be outdated if someone downloaded the old embeds. be my guest
>>
File: 7105137.png (858 KB, 3220x2073)
858 KB
858 KB PNG
I'm a retard and didn't think the CYOA theme would change all unset fields in the default ST themes even after going back. Because of course it remembers last used settings, and does NOT save everything in the default theme—just the changed fields. It's just like the proxy-swapping fucking up presets.
Is there any way to reset a theme to defaults? I miss wider chat.
>>
File: Spoiler Image (277 KB, 652x732)
277 KB
277 KB PNG
>>41665835
>>
File: look here dumbass.png (44 KB, 342x429)
44 KB
44 KB PNG
>>41665938
1. You're a retard it says to back up your shit.
2. see pic
>>
>>41665938
Try importing this? https://files.catbox.moe/5e06ns.json
Should be the default theme.
There also https://rentry.org/meta_rice_list if you want more themes to test.
>>
>>41665963
I won't dispute the first part, but in my defense what happened was that I checked if it's available in case something doesn't work. I made sure it was:
https://sillytavernai.com/download/ross-v2-default-sillytavern-theme/
It just that this doesn't work. I think the themes must not contain info about most of the unchanged fields for some reason. When a different preset overwrites a bunch of them, swapping to the fresh defaults only changes some of the setting back.

>>41665969
Thanks, this works.
>>
File: screenshot.png (61 KB, 877x332)
61 KB
61 KB PNG
>>
>>41665256
not the lore book, the char book, the place for extra char info , char personality and example of speech pattern
>>
File: Gemini.png (21 KB, 565x260)
21 KB
21 KB PNG
WTF. Already?
>>
>>41666162
Wait, what the fuck. Didn't they release a model a couple weeks ago?
>>
>>41666212
Yes. 1114.
>>
File: gemini_1121.png (153 KB, 839x720)
153 KB
153 KB PNG
>>41666221
they also released learnlm-1.5-pro-experimental model two days ago. so three models in one week. that's an academy record!
picrel, example on new Gemini 1121 . it writes like 1114 imho but follows commands better. censorship is the same (none). I guess 1114 is their "loose" model but 1121 is "follow commands strictly" model? dunno. they feel the same for me
>inb4 preset
https://rentry.org/x3ywiub6
>>
>>41665901
Both Posey and Mistress Moonbow should have the new lorebook now.
>>
is there any jb for gemini? or do the chorbo ones work?
>>
>>41666385
It seems to work fine with the CYOARPG JB with the occasional error. The only thing I found is that it needs to have a message to go off of so greeting are mandatory.
>>
>>41666389
Great, I was having fun with that one
But how's it doing with the hardcore shit?
>>
>>41663407
>>41663335
Oh my god what JB is this, or what extra tweaks are you doing? It's genuinely putting a smile on my face, it's so goofy.
>>
>>41666408
The JB is here >>41663387; I use it with the new Chorbo.
It’s the one with the dice roll that’s often posted here with a panel for the "Naigga Cyber Pratchett" prompt.
While bad for the filter and NSFW, I find the new Chorbo really good for SFW. Way more creative, proactive, and sometimes less positivity-pilled.
It works very well for silly prompts with fun cards like this Cadance or a rewrite I made of Lesson Zero Twilight greeting. In that one, for instance, the model has way fewer problems with Twilight using her magic on me and doing fun, crazy shit.
I need to test it more for that, but I believe it could be good for adventure too.
>>
>>41666494
whats chorbo?
>>
>>41666526
Chatgpt-4o-latest=Latte=Chorbo
>>
>>41666403
I mean I was using it for some
>rape
but I haven't really tested it with much else. From my past experience Gemini isn't filtered much unless its filly.
>>
bros how do i fix being addicted to AI mares
>>
>>41666551
Do something else in parallel. Just try something if you’re unsure what to do.
Also, as much as I like these models, their limitations can make it easy to feel 'chatbot fatigue' after a while, so taking breaks to do other stuff is always a good idea.
>>
>>41666377
Nice.
Only 2 cards using the book? Clearly we need moar.
>>
>>41666551
just take a break, touch grass, do something with your life, anything in moderation, they're not real, i view this more as a visual novel game than anything else. its never too late to snap out of it and get a healthy headspace again. remember, too much boners cause your brain to shrink, and its not a joke.
>>
>>41666579
I'm not a big G5 fan, but the Posey card is fun.
>>
>>41666569
thanks anon, ill prob just temp delete sillytavern my life recently has become
>work
>gym
>mares
>sleep
>repeat
>>
>>41666600
At least you were doing other stuff, like going to the gym. But yeah, if you feel like it’s taking up too much of your life, just take a long break.
>>
File: SOMNAMBULANT SAVAGERY.png (153 KB, 931x591)
153 KB
153 KB PNG
>>
>>41666585
Love me some Poser, what would some other fun5 cards be. Sat AM villain time with the QUEEN?
>>
>>41666708
I would play that.
>>
Based thread, haven't been here in a while, I assume free AI mares is still difficult, but which is the best model nowadays? and which look promising for the future?
>>
Is there any way to refresh IP count for Column? I fucked it up by using on both agnai and risu. Now I'm almost running out.
>>
>>41666890
Outside of emailing the owner, probably not.
>>
File: scared.gif (623 KB, 253x272)
623 KB
623 KB GIF
this thread is bad for my wallet, every time I browse it pushes me closer to getting another 3090 to run slightly bigger models locally
>>
>>41667067
It's a one time spend... An investment, if you will.
>>
>>41667052
The emails on the other rentry I assume?
>>
does that cyoa work on any other models?
>>
>>41667104
Only one way to find out Anon.
>>
>>41666850
Claude is rare now. So it's a new model of GPT called Chorbo or Gemini, I would say.
>>
>>41664043
Does this mean there’s a chance at chorbo being fixed and brought back to normal? God I hope so.
>>
>>41667104
Kinda works on the latest gemini and GPT. Breaks on mistral.
>>
>>41667418
>I miss my wife Claude, Tails.
>I miss her a lot.
>>
File: HA.png (1.18 MB, 2023x2023)
1.18 MB
1.18 MB PNG
>>41667482
I understood that reference.
>>
>>41667067
Just buy openrouter credits dummy
>>
File: YesIAm.png (9 KB, 691x24)
9 KB
9 KB PNG
Cute.
>>
>>41667482
Alfred was the best only good part about Trannycube's fan dubs.
>>
MM is kill.
>>
>>41667622
he's just adding his 2000 anthropic keys.
We will be so back you'd think we never left.
>>
>>41667599

Super glad he's been pretty much absorbed into Lythero's coterie so he doesn't have to slum it on that sinking ship. Can't think of a single bit of content they released that was even half as good as their dub of SA2's Evil/Final Route... the part with the most Eggman in it, what a coincidence.
>>
File: Dumb.png (247 KB, 959x651)
247 KB
247 KB PNG
The ultimate defense.
>>
>>41663144
card looks more in-depth than estrus sim 1. any chance of a fork for estrus sim 2, but for g4?
>>
>>41664244
>What's the max context for MM's Gemini?
Speaking of, can we expect an increase there anytime soon?
Soul is finally back on the menu ever since the new Gemini release.
>>
>day 4 of Column being fucked
>>
>>41667875
What proxies have the new gemini and how do I use it? I don’t know if I’m using the model name wrong or not, but it just seems to get stuck in infinite loading for me. It’ll some times give some error too.
>>
>>41667947
MM has it, try the latest experimental snapshot and Avani's JB.
>>
File: Kino.png (78 KB, 807x211)
78 KB
78 KB PNG
>>
>>41667890
It was working for me most of last night, though it's been actually busted since like 6 ish hours ago, to my knowledge at least, I was sleeping for most of that
>>
>>41667067
How many 3090s are you at right now?
>>
>Column dead
>MM dead
OH FUCK OFF
>>
>>41668064
>>
>>41667963
Since MM is currently dead, is there anywhere else that's got it?
>>
>>41667947
Unreliable has it
>>
>>41668103
For me Gemini Exp on column just gives blank responses. Gemini 1.0 Pro is the most recent one that works on my end.
>>
>>41668102
wulfs, easy archive search
>>
I had a feverish dream today, in it our bug opened a bugproxy and the main condition for getting in was to behave like a decent person in 'cord for a month.
>>
>>41668161
>behave like a decent person in 'cord for a month.
That's a nightmare.
>>
File: SweetCelestia.png (516 KB, 605x600)
516 KB
516 KB PNG
>>41668161
The nightmare timeline came earlier than expected.
>>
>column
I was pretty surprised to see myself in the top twenty. I thought I was using it sparingly.
>>
Is there any spoonfed guide about lorebooks and world info? I've never used it, but now I'm testing the cyoa jb and I think it would be interesting to put more background about the world, but there are many options and it's quite confusing
>>
>>41668281
https://rentry.org/AdvancedCardWritingTricks#character-lorebook
>>
i can't believe i'm in the top 10 for column i'm so sorry
>>
File: 1712626539682433.gif (609 KB, 320x233)
609 KB
609 KB GIF
>>41668287
Thank you
>>
>>41668240
>>41668295
Kek, I'm just about the midpoint (a little over it) and I thought I used it a lot.
But also, Pareto's principle. I threw the stats into an excel sheet and apparently the top 20% (30 locusts) are responsible for 83% of the usage.
>>
>>41668326
I wish I could check where I am but I unfortunately don't have access to my own token, unless there's a way to extract it from my preset somehow.
>>
File: file.png (18 KB, 963x79)
18 KB
18 KB PNG
>>41668393
?
If you have it saved in one of your presets, you should be able to go to the second tab and see your token.
>>
>>41668415
Agnai.
>>
>>41668240
Huh. There's actually a lot more people using column than I thought.
>>
>>41668620
Yeah, 186 a lot.
>>
>>
>>41668760
>>
>>41668760
But that's Fluttershy.
>>
>>41668780
It's clearly Rainbow Dash. See her horn? Pegasus.
>>
>>41668780
I've redesigned the card a bit. Now instead of Fluttershy guessing my fetish, there's Applejack trying to sell me apples or other farm crap.
>>
MM is back up
>>
>>41668871
Nice. Same URL too.
>>
mares are expensive
>>
>>41669120
Claude on OpenRouter?
If so, did you try https://rentry.org/pay-the-piper-less?
>>
>>41669127
Intresting. I never have really messed with OpenRouter. I've just been using GPT4
>>
>>41669326
>GPT4
just use a proxy, anon...
>>
>>41669326
Lol wow...
That / models better than that are free / nearly free everywhere.
>>
>>41669326
there are better options, anon. read the newbie guide in the OP all the way till the end, it’s worth it in your case.
>>
>>41658102
Ideas? I could try dot, I guess.
>>
>>41669770
I mean other then just picking a character and giving it a go.
-Celestia/Twilight organizes a job fair to try and help. (You) are dragged along or help organize.
-Checking over the OCs it seems most don't have cutie marks, could do something around that.
- Going more for the misfits side you could do a card around the mane 6 having the wrong cutie marks.
>>
>>41669336
>>41669492
Are there? I read the wiki and used the table from there and the one from /aicg/ and GPT 4 was fairly good all things considered. As for the proxy business I hadn't given it much thought because from what I understand the price doesn't change. But if I'm being a retarded newfag please feel free to correct me.
>>
It looks all the proxies for latte are being locked away though a discord so now not only is the model fucked I can't even mess with and try to make it work.

Do I try to get into this aicg discord or what? Man this is so frustrating.
>>
>>41670216
aware aggregate c pick
>>
>>41670216
Get in my Discord and you can have all the hot horse smut you want, hand-typed.
>>
>>41670222
what
>>41670233
I'm willing to go into a discord if it is an escape from the shitshow that is 4chan. At least I think that's the impression I'm getting is people are fed up with aicg and I don't blame them.
>>
>>41666890
>>41667052
>>41667102
Can I really reset IPs if I sent email?
>>
>>41670261
The owner is pretty reasonable (and surprisingly kind), so it's worth a shot. Good luck.
>>
>>41670267
Thanks. I'll try.
>>
File: 1637120062759.png (212 KB, 676x700)
212 KB
212 KB PNG
And speaking of column... I know at least one of you managed to get it running on Risu.
If it's not too much trouble, could you please write a little guide about how you did it? I'm too retarded to figure it out on my own.
>>
>>41670216
Anon... This is very, very obvious bait that’s been going on for a lot of threads. They don’t have proxies.
Don't bring this shit here.
>>
>>41670441
So the proxies are just gone? If not there then where.

Regardless just testing it with a alt and call me a pussy but I'm not liking what I saw at first glance. I'm a massive faggot and I rather be depressed with my inability to chat with ponies then hanging out with people I clearly wouldn't get along with.
>>
I'm just so confused how the discord is bait stuff when
https://rentry.org/chorboproxies
Is all locked up with the only link working seemingly to be a fake "gotch your IP' scam
>>
Really enjoying this CYOA thing, it works great with a lot of pre-made cards that are RPG focused too. And adding ponies is always so easy. I'm using a silly Isekai card to send some humans to G1 Dreamland and D&D fight Tirrac and the Stratodons, it's some good whimsy.
>>
>>41670687
How is the model doing with G1?
>>
>>41670698
Using ChatGPT 4
It instinctively knew a few G1 ponies though it doesn't seem to be 100% on their descriptions, but it seems to have personalities worked out
I'm giving it hints and instructions about stuff but it didn't need much more than "Do MLP G1" to pull this out
>>
>>41670702
O and here's the card

https://files.catbox.moe/bnpbhx.png
>>
File: latest.png (5 KB, 610x50)
5 KB
5 KB PNG
>>41670702
I’d try this version in Tavern. It has a much nicer writing style, IMO.
>>
>>41670731
Ooh you're right, thank you
>>
>>41670784
Yeah this model is really good, I like this style way more! Thanks anon
>>
>Find latina pinkie pie bot
>It's all fucking Spanish
>The only non Spanish one is anthro that wasn't given any Spanish flavor phrases
A sad night to be sure
>>
>>41670930
just fucking tell a different english speaking card to translate it for you 4head
>>
> "message": "Cannot read properties of undefined (reading '0')"
Anyone else getting this error with Gemini?
>>
>>41670962
Using Google AI studio as the Chat Completion Source or OpenAI with external models checked?
I would try both to see.
>>
>>41670949
Ai doesn't translate the intricoes or slang of languages, it gives a best guess for what it thinks something means in context and over half the time that shit sounds stale or just incorrect
>>
getting blank responses on column
>>
>>41671062
I turned off streaming and it seems to work fine now
It stutters a lot with streaming on in the first place so might be something to do with that
>>
>>41671102
Streaming off just gives me the something strange and unknowable error.
>>
>>41671101
We're on Source 2
>>
>>41671295
but it says src 1!
>>
>>41671340
Not anymore it doesn't
>>
File: Cute.png (82 KB, 848x192)
82 KB
82 KB PNG
>>
Oh look a new proxy, I can't to do something later tonight-
>Dead
Okay some people have some explaining to do. "The discord is bait don't bring in this thread" my actual dirty fucking ass. It absolutely seems to be the only source of chorbo proxies as all the others are being deleted. I don't even understand fucking why. The filter is fucking it up anyways what's the fucking point in making it so hard to actually get in. All I want to do is actually test people's claims of knotty or Avani being better at getting past the filter because Smily just doesn't work but apparently if your not sharing logs on the cord then you're not worthy of having proxy links. So insufferable.
>>
>>41671691
in OP there is a novice guide
check MLP stuff at the bottom
>All I want to do is actually test people's claims of knotty or Avani being better at getting past the filter
it is not. and for now Avani took a break from Latte. she is more interested in Gemini models which are easier to wrangle. to make things worse, OpenAI broke (or deliberately changed) how samplers (temp/top_p) + logit_bias work. which is infuriating that model acts fine on [temp 1 / top_p 1] but it you do [temp 1.1 / top_p 1] it goes completely nuts and becomes unusable. so you tone down sampler a bit to [temp 1.1 / top_p 0.95] and then model writes the same as on default [temp 1 / top_p 1], so... what's the point? and if you set top_p to anything but 0.99-1.0 then all logit bias go to hell. so yeah she is fed up with Latte and 1120 for now. as for other JBs - Knotty's JB and Karu's Cheesecake still can do NSFW occasionally
>The discord is bait don't bring in this thread
there are lots of closed or privated cords and none of them are doing "join us and we will give you a proxy or super prompt to let you coom for free :333" - stuff like this is a bait
anons there do not like talking about discord because it brings drama and thread is getting shitted fast. noone likes namefagging and attention-whoring
yes it is true that lots of anon there (counting me) are using discord. but we do not post stuff from servers in 4chan threads, unless it is something important like maybe proxy updates. the more people bring up discord the more it prone to become "look what this local e-celeb botmakie said about this e-celeb promptmakie - omg - gossip-gossip-gossip", fuck this. we have twitter for this sodomy. if anons want to see cord screencap - they would go to cords to read them
>https://rentry.org/chorboproxies
two proxies at top are long-standing proxies providing Gemini and GPT. but people started actively forcing them causing >rape of the keys. authors were forced to put a discord membership (be in Scylla discord server for at least 2 week) as some sort of leverage to sustain/prolong the keys. it is what it is, Anon
>>
>>41671691
>"The discord is bait don't bring in this thread" my actual dirty fucking ass.
1- The one spammed in the other generals are bait. All the "join us and you will have a proxy!" stuff is bait.
2- Not bringing Dicksword into a fucking 4chan thread is just common sense.
>It absolutely seems to be the only source of Chorbo proxies as all the others are being deleted.
You posted a rentry here—just go to the last line and read what’s in parentheses. We also often recommend newfags to read a certain guide.
>The filter is fucking it up anyways what's the fucking point in making it so hard to actually get in.
It’s limited resources, and for public proxies, we depend on other anons who decide to invest their time to make them run. Big surprise, it has its limits. The filter has nothing to do with it. Sometimes there are droughts, and sometimes you’ll have a day without a proxy. That's also why they often can't just be easily displayer or advertised. That’s just how it is.

>some people have some explaining to do
>my actual dirty fucking ass
Now, I’m normally a chill motherfucker, but fuck, did your post rub me the wrong way, so:

Most of your posts makes you look like a really whiny, self-centered faggot. I don’t think I’ve ever seen you contribute anything positive to the thread. You only take and complain about it. I hope you can see that.
>>
>>41671770
>Most of your posts makes you look like a really whiny, self-centered faggot. I don’t think I’ve ever seen you contribute anything positive to the thread. You only take and complain about it. I hope you can see that.
The very nature of 4chan is being anonymous so how are your able to "tell" I've contributed nothing positive. Sure I'm bitching like a faggot now because it feels like all my questions have been bushed off or ignored. Like no one actually wants to get people in and only wants to gatekeep to make sure there usage of the ai isn't fucked over by new comers. It's extremely frustrating and acting like a bitch seems to be the only way to get people to fucking engage. Hell it worked again and actually got some answers finally. If the process wasn't so frustrating and I could reliably work with the AI I would then be able to contribute even more than the limited amount I've been able to do.
>>
Only GPT4 latest got hit or it was a every model?
Seems like every model is acting different and mostly worse despite the JBs
>>
>>41671878
Latest is the only GPT model that is actively updated without versioning. The other models are snapshots that aren’t modified once released.
>>
>>41670930
>>41671066
Is it in the bot list that we've got here? I could give it a look and try to Spanglishify it or whatever you're hopin for.
>>
>>41671913
NTA but I think this is the one.
https://www.characterhub.org/characters/Anonymous/pinkie-pie-latina-d505c190fd32
>>
>>41672063
>Bringing RL shit into pony
go die in a hole
>>
>>41672076
Okay now tell that to everyone who inserts as Anon.
>>
>>41672086
"Anon" is not real world baggage.
>>
>>41672086
Anon is a fantasy thought construct so far removed from reality he might as well be a new race.
>>
File: file.png (17 KB, 820x37)
17 KB
17 KB PNG
I don’t know if it’s just me, but Chorbo went schizo for a little while on every output before returning to normal.
I’m using a normal preset, and my penalties are set to zero, so it’s not my settings. It was also brief and went back to normal shortly after.
>>
>>41672150
not right now but yeah GPT sometimes does this. probably GPU cache is going insane
>>
Feels like im about to blog post again as my response before was still admittedly heated. I hope that's okay
>>
>>41672203
Don’t blogpost. That’ll only bring drama and off-topic discussion. Just move on.
>>
>>41672214
I'm trying in my own way to help out but I feel I need to speak clearly for a second.
>>
>>41672223
Then speak clearly but don't make it overly wordy or emotional. Share your thoughts but y'know, don't be a drama queen. Or well, share them if their related to bots and AI and stuff. If it's totally unrelated then yeah nobody cares
>>
I was overemotional and gave off the wrong impression that I'm a self centered asshat. My approach to getting usage of AI will eventually be something I can pay forward to others as I feel there are other lurker only users like me that would appreciate it and I'm willing to push though my social anxiety to share logs, cards and the like (with info on how best to that as a mobile user)

What direction do I need to to get access to a actual real discord (though I still don't fully by other being bait and think there is some miscommunication going on or I'm just THAT gullible which is entirely possible)

Blah blah I'm thankful for those that get the keys, you do a thankless job and I genuinely appreciate those who help out the rest of us.

My comment summarized as succinctly as I could, I hope.
>>
the funny thing is that i was able to get a permanent key just by being really nice and asking politely

you catch more flies with honey than with vinegar, anon
>>
>>41672269
Once again. Read the novice to advanced guide, all the way to the end. You've been led to the water multiple times, but you still refuse to drink.
>>
>>41672269
You need to look for something that will provide you a small quest. Complete that quest and you're in. It's that simple. Good luck. If you can't do the quest, you don't deserve the key.
>>
Stop engaging with him. They'll go away soon once they realize no one cares.
>>
>>41672294
>>41672298
Thank you. I'll try my best.
>>41672300
Fuck you
>>
File: 1732389410807700.gif (1.28 MB, 500x281)
1.28 MB
1.28 MB GIF
>>41672269
>>
>>41672416
off topic but there are no words to describe how much I hate rainbow dash
>>
>>41672416
It was a lot longer and more cringe before trust me
>>41672429
Why don't you hatefuck her like the user mentioned with Rarity a couple threads ago.
>>
Bros, I claimed the google cloud platform free trial but I have no idea how to set it up on sillytavern, could someone help me?
Sorry I'm a techlet
>>
>>41672530
I don’t think many have tried it here since you now need to enter your credit card.
But there was a guide posted for it—you’ll need to host your own proxy locally.
https://web.archive.org/web/20240721035801/https://rentry.org/gcptrial
https://github.com/cg-dot/oai-reverse-proxy/blob/main/docs/gcp-configuration.md
They should still work, but I'm not sure if they still allow Opus.
>>
File: NiceTitles.png (87 KB, 883x287)
87 KB
87 KB PNG
>>41672150
Happened again, and now I’ll never know which mare it could be...
>>
>>41672815
are you sure it is not your preset? because the timing of this ""glitch"" vs narrative is just too convenient. maybe you have something like "keep suspense up" or "add cliffhanger"?
>>
>>41672544
>They should still work, but I'm not sure if they still allow Opus.
Apparently, you need to add 8 bucks to your account before they will let you use it.
>>
>>41672827
I don’t think so. At one point, it even started talking to me in Arabic with a card where it didn’t make sense. It does this for a little while, with all swipes being fucked up, and then it goes back to normal.
The preset is a very basic one I’ve been using for days with a classic 1,0,0,1 setup. I only started having this issue when I first posted about it.
I’ll install a fresh Tavern version and see.
>>
>>41672870
hm. then it seems something is off with model. I wrote a bit earlier about logit_bias + samplers going haywire >>41671755 - maybe that is your case as well
>>
>>41672905
I do use logit bias. I will test without.
>>
>>41672544
Ok I did all the steps and it seems to connect the proxy with st, but when I type a prompt it gives me a blank or an error that says there are too many concurrent requests.
I look at the proxy and it says there is 1 proomptersInQueue and the estimated time goes up, at first I panicked and thought someone stole my key or something, but I close st and the queue goes back to 0 and no wait so...it's just me
Any idea what the fuck is going on?
>>
column is on maintenance?
>>
>>41672995
>(status?)
>trying to fix system prompts
>>
File: 5y54u75ui6i56i46i.png (607 KB, 946x628)
607 KB
607 KB PNG
she know what's best
>>
File: Cadance_bump.png (172 KB, 898x601)
172 KB
172 KB PNG
>>
File: mlp Celestia 0.jpg (2.63 MB, 2700x3000)
2.63 MB
2.63 MB JPG
Retard here, been struggling my way through this, never used Github or AI or anything. So far managed to download GitHub, and Node.js, and SillyTavern, and got it to open, and now I'm exhuasted and ready to sleep.
Hoping for answers tomorrow.
So, on Connection Profile, do I need to download a particular API, whatever that is, or select one of the available options?
And how do I download a Ponydex character, or do I need to link the web page somewhere?
>>
>>41673504
Depends on the API. If you wanted to do GPT 4 you would click under "OpenAI Model" and select it. You don't "download" a particular API you get a key from the respective website. As for Ponydex use the "mlpchag.neocities.org" and just find a card you like and download it. Then upload it (use Sillytavern manual.) I'd suggest reading the guide at the top. Hope that gave you an idea on how to fix your problem. Have fun with your mares anon <3
>>
>>41673504
https://rentry.org/lunarmodelexperiments

here bro, i made a whole tutorial on every single aspect of setting it up for local, if you plan on using a proxy, most of the information in this tutorial is still relevant such as character and setting up silly tavern, let me know how it went
>>
>>41673504
>And how do I download a Ponydex character, or do I need to link the web page somewhere?
The metadata is in the image, so you can just use the image itself.
Drag and drop.
>So, on Connection Profile, do I need to download a particular API, whatever that is, or select one of the available options?
For local models, we have a guide in the OP.
For non-local, read the novice guide all the way to the end. It's worth.
>>
>>41658075
>>41670930

Here it is. Not super high effort, honestly half of the original Spanish text was redundant Latina obsessing, coomerslop notes, or what seemed like copy pastes from Pinkie's wiki page or whatever so I tried to make it more concise. There wasn't even any Spanish example text or anything, so I just kinda specified to make her speak Spanglish. I could redo it as just a word for word translation or maybe try to emphasize the accent more if you'd prefer but I'd like to think this is good enough. Kept a little bit of the coomerslop that seemed like it'd at least lean into the gimmick though.

https://files.catbox.moe/4ss81e.png
>>
>>41673678
Oh right, also changed the picture but obviously you can just put the other one back on, I just prefer that not being the first thing I see on a card.
>>
>>41673020
>"By the time I'm done with you, anon, you'll be the best pony lover in all of Equestria. Mark my words!"
>t. Pinkie Pie (The rapist)
>>
>>41673678
>>41673680
Thank you, I wanna say I did try to find a half assed translating bot like the other guy suggested and it went as expected it either ignored parts or just up and changed some details as it pleased
>>
File: 2UEk3QZ.jpg (19 KB, 527x612)
19 KB
19 KB JPG
>set the model to GPT4o
>load up corpsefucker
>slowburn through 30 messages
>hey this is kind of soulful and fun
>finally get down to flavor town
>sorry, I can't assist with that
Is Gemini 1.5 Pro uncensored and unhinged if I get an API key direct from Google? I tried using Gemini Pro through OpenRouter, but it has given me refusals. We live in the gayest timeline. All I hope for right now is to roleplay my steamy fetish fantasies with a chatbot. If I die in my sleep tonight, I can accept that. Every day brings my closer to ending it all in this hopeless world. I just want to make love to AI mares, yet some patronizing dickhead in his office decided that I should not.
>>
>>41673893
If you’re doing Slowburn, you can start with chatgpt-4o-latest and then switch to gpt-4o-2024-08-06 with a JB from https://rentry.org/jb-listing#gpt-4o for sexo, like Camicle.
Or switch to Gemini, yeah. Gemini-exp-1114 or Gemini-exp-1121 are the favorites. Never got filtered on those. Depending on the proxy you’re using, you might have access to it.
>>
File: Blue.png (189 KB, 865x548)
189 KB
189 KB PNG
>>41672447
>hatefuck her like the user mentioned with Rarity a couple threads ago.
I actually love Rarity; I’m just a sucker for stories where Anon and her start off absolutely hating each other, only for it to evolve.
It’s fun, whether it’s for slowburn or just smut. For adventure, you could strand them somewhere like the Everfree and have them try to survive together.
>>
>>41673497
Out of interest. Is that smiley jb? I've noticed I get a lot of replies where the mare repeats something I've said. I don't know if it's a gpt thing or a smiley thing.
>>
Avani's Gemini JB is really good, but it keeps trying to write entire messages for other characters in group chats.
Does anybody have any solutions?
>>
>>41674066
Don't use Avani.
>>
>>41674068
Are there any other good Gemini JBs already?
>>
>>41674066
I’ve never used Avani or Gemini for group chat, but presets often include instructions saying they can talk for {{char}}, the world, and NPCs around—just not {{user}}.
So I guess for it, it’s fine to talk for the other NPC/bot characters.
Try adding somewhere that it should focus mainly or only on {{char}}, preferably as low as possible in your preset.
>>
>>41674078
>Try adding somewhere that it should focus mainly or only on {{char}}, preferably as low as possible in your preset.
I tried doing that in the perspective options, results not encouraging. But I just saw that there's an update for the new gemini so I'll try that instead.
>>
>>41674083
>>41674078
The new version actually has an option for group chats lel
>>
After a long bount of trail and error, I have a semi functional Paprika to throw at any MLP bot I desire. On that note, I need an opinion, should I give her a ponut? If so how would you properly describe one and should I segment it separately from the other sexual details section I made?
>>
>>41674104
I usually avoid NSFW details in cards that can be both SFW and NSFW, as it could make some models unnecessarily horny, and users might use a preset for detailed equine genitalia.
But the meta and models used have changed a bit since then.
Maybe consider a pure SFW version and a fork with NSFW details?
As for the ponut, you can try describing the shape, etc., but from what I’ve seen, unless you ask the models to focus on the differences, it will mostly be treated as 'normal.'
>>
>>41674032
No this was the avani for 1121 but gemini was shitting the bed so I swapped it over to GPT.
I've never gotten an AI to never repeat what I say regardless of model or JB but you can minimize it.
>>
Still having fun with my lil story, GPT 4 latest is way better than the other ones I was trying. We G1 baby
>>
>>41674505
For SFW stories and prose, I still find Latest to be very good. Also, if I’m not mistaken and you’re the anon who pays for the API, I’d recommend the same reading material that’s been mentioned several times in the thread.
>>
>>41674541
nup i'm fluttershy! but tysm
>>
File: file.png (2 KB, 200x200)
2 KB
2 KB PNG
>>
>>41674575
for some reason, this makes me want to play Super Nintendo
>>
>>41674575
Testing it.
It’s working for me, but if I have streaming on, the answer gets cut. I need to disable it to get a complete answer.
It’s very slow too, but definitely better than nothing. Worse than Source 1 but better than 2 (I could never get anything to work with 2).
>>
>>41674505
unironically I think it could be a replacement to sonnet if it were not for the filters
>>
>>41674575
WHERE are you getting this from? it still says src 1 for me
>>
>>41674620
ctrl-f5
>>
>>41674620
CTRL+F5
>>
>>41674605
everything results in error 500 for me (something strange and unknowable)
tried both streaming off and on
>>
>>41674622
>>41674623
mobile
>>
>>41674629
if you're willing, could you send more details, including a full copy paste of your terminal output to faggotproxy.elusive847@passinbox.com?
>>
>>41674634
(to clarify: you don't have to send your current terminal output or anything before the request that errors, just the full output starting from a request that causes the error. swipe on a blank card or some such)
>>
File: nopus.png (16 KB, 640x218)
16 KB
16 KB PNG
>>41672972
i tried getting the gcp trial to work a while ago and the problem is this. i think google gimped it at some point. still don't know how to fix it.
>>
>>41674670
you need to fill your gcp account with around $10
>>
What's Oobabooga like? Is it good? Sexual limits?
>>
>>41674715
It's a front-end to run local models. Not a model.
>>
Is gemini 1121 suddenly not working for anyone else that uses it? I’m just getting infinite loading.
>>
File: jeet hate.png (15 KB, 1409x298)
15 KB
15 KB PNG
>>41674670
yeah, I tried to request a raise to 50 and they emailed me this
>>41674675
Is there any way to see what's the minimum amount to be added?

Well I am not in a hurry I will wait those 2 days and request again, if they send me the same I guess it will be time to open the wallet....
>>
>>41674797
check - works on my key. so if it doesn't work then high chance it is a proxy issue. maybe keys on proxy are ratelimited?
>>
>>41674811
>Is there any way to see what's the minimum amount to be added?
From what I've seen, people keep repeating that $8 is the lowest you can go and still get it to work.
>>
>>41674797
happened to me yesteday, but today it works. maybe try restarting silly or something?
>>
Hello, someone posted the novice-to-advanced guide on another board, and I just wanted to say thanks. I got started with his hobby a couple of months ago, but even so I still lack knowledge in many areas of how the AI works and what models are recommended, so I wanted to say that it's a pretty cool guide.

With that said, I have a question that maybe someone here can help me answer. Couple of days ago I noticed while using Chorbo that it was filtering way more, sometimes not even letting my prompts go through, and read there was apparently an update in the filters? So I wanted to ask if there's any JB that can already pass the filters or if I gotta stick to SFW stuff for now?
>>
>>41674541
Yo that was me. I still don't understand what you mean. Are you suggesting that there are cheaper ways for GPT-4 or that there are better alternative?.
>>
>>41674902
Who's your favourite pony?
>>
>>41674920
It's been years, but back when I watched I think my favourites were Applejack, Twilight Sparkle and Celestia
>>
>>41674902
>on another board
/vg/?
>So I wanted to ask if there's any JB that can already pass the filters
the thing about Chorbo is that developers still been tweaking both this model and new 1120 snapshot, we are waiting until they stop already and leave them the way they are
for JB - Knotty's JB can get some NSFW (mostly consensual and wholesome). there is <del>-related prompt, enable it
Karu's Cheesecake is successful sometimes too
you can see both them there - https://rentry.org/lattejb
other JB may work as well. it depends mostly on what your content is
if not Chorbo then check previous 4o models - 0513 and 0806, they are less filtered but more dumb. do Chorbo for as long as you can, switch to them for sex scene then back to Chorbo. alternatively switch to Gemini. JB for Chorbo will eventually come. it is like for now OpenAI are playing whack-a-mole so it is pointless to JB it for now
>thanks
<3
>Applejack, Twilight Sparkle and Celestia
based taste
>>
>>41674938
>/vg/?
No, /vt/
>the thing about Chorbo is that developers still been tweaking both this model
I see, so once they stop tweaking the model then the JB's will be able to catch up right?
Also I have been using Smiley as my JB and it used to do NSFW almost always, So I guess I'll try another JB or the snapshot method
>based taste
Thanks, I remember Celestia showed up so little in the early seasons I depended completely on fan artists and fanfics to get my fill of her. Even so I still loved her design
>>
>>41674912
Both. Go read, man.
>>
>>41674964
>vt
wAIfu thread my beloved. I had no idea that this thread still exists. good to see it is still going on!
>>41674964
>JB's will be able to catch up right
yeah, and also in last 2 days they toned down filter a bit. seems to be some users (not coomers) complained about too sensitive filter. so lets wait a bit. depending on what you want and what you are doing in story - Chorbo may already work fine for you with some JB. it is very a trial-n-error thing
>>
File: Mare.jpg (213 KB, 1500x1500)
213 KB
213 KB JPG
Just saw that the Chatbot Arena, previously LMSYS, is now
https://lmarena.ai/
Further emphasizing the importance of mares in AI.
>>
>>41674938
>>41674964
>for JB - Knotty's JB can get some NSFW (mostly consensual and wholesome
As someone who was also primary using Smiley for Chorbo and basically had everything get filtered I can give attest that the updated Knotty has managed to break through 90% of my Smiley chats, and I'm stuff is definitely not wholesome and is non-consensual. It was mostly just continuing previous chats mind you but I did test a wide variety of them, from just starting, to almost over and in the middle.

The only thing Knotty got filtered was with a scenario card strangely enough, and I think the issue seems from me having a large amount of the original story (as the scenario is meant to try and help create a what if) posted in the chat because I didn't understand that 128k tokens doesn't actually mean the ai will handle that much in chat history. I still managed to get responses eventually but I need to rethink the entire card as a whole.
>>
>>41674999
>good to see it is still going on!
It is! However the latest Claude drought has hit it pretty hard, however I still hang out there, it's a great thread
>>41675017
Guess it's worth a try, BTW how is it for SFW stuff? does it stick to that or does it try to derail towards NSFW often?
>>
>>41674632
Firefox? Tap the icon next to the url then clear cache. Other browsers should be similar.
>>
>>41674629
The issue has been resolved, or at least on my end it has
>>
>>41674968
it burns
>>
>>41675070
Though I have also discovered the censors are higher on this source even on Sonnet
>>
>>41675048
>BTW how is it for SFW stuff
Don't do sfw so don't know
>>
If I pay for Claude will I get banned for doing smut? I don't give a shit about being logged, but I don't want to waste my money.
>>
>>41675080
The easier way is through OpenRouter, and they never ban. I heard Anthropic can ban direct access API, but it's rare and mostly for cunny from what I've seen.
>>
>>41675090
>but it's rare and mostly for cunny from what I've seen.
No. They go by waves and hit everyone that has gone over a certain threshold of nsfw prompts vs sfw prompts.
>>
>>41674859
I’m using risu and it was working just fine earlier this morning.
>>
>>41675070
Broken on Agnai.
>>
>>41658075
Dunno if anyone is interested in some autistic power fantasy in their pony chats, but here you go:
https://www.characterhub.org/characters/Kadon/wardens-of-harmony-cf750e01d9b6

Here's a card that lets you become a magically enhanced, monster slaying badass in Equestria. The Wardens are a headcanon I've been working on for a long time, but for the purposes of a character card, I've cut down a LOT of my lore tism and left only the important bits.
In the first greeting you're a fresh trainee, just having passed a magical aptitude test and about to start training with your new partner.
In the second, you're a fully fledged Warden, one of the most respected and powerful operatives of Equestria and you're just coming in for a new mission briefing.
If anyone actually ends up playing this, please post logs of your adventures. That'd make my day.
>>
>>41675074
Alright, just tried Knotty and when I did so I remembered I had actually tried it some time ago too.

I found it's very good for NSFW but I'm not entirely satisfied with it's SFW scenes. I think I'll alter between Smiley for SFW and Knotty for NSFW (That's possible, right?)
>>
>>41675265
yes very possible, go ahead!
>>
File: Bots.png (69 KB, 458x284)
69 KB
69 KB PNG
>>41673678
Added, thanks anon.
https://mlpchag.neocities.org/view?card=Anonymous/Latina%20Pinkie%20Pie.png
>>41675178
Also added, thanks a lot.
https://mlpchag.neocities.org/view?card=Kadon/The%20Wardens%20of%20Harmony.png
>>
>>41675270
Thanks! last question and I'll stop for now: the Knotty JB that I have to download is the one that's on this Rentry right? https://rentry.org/lattejb. If so, whata re the other links for? (Regex, thinking, etc)
>>
Okay column's up but it's really fucking weird. I tried to prompt it and it made one big frankenmessage where it also spoke for my persona
>>
>>41675119
yeah, it's nor working for me too now.
>>
>>41675324
>right
yes!
>If so, whata re the other links for?
they all are regex files. regex is a text-processing syntax to manipulate text strings. lots of programing languages and OS support it. including software like notepad++ or SillyTavern. those extra files on Knotty page can be installed into SillyTavern and they will process text generated by model (or you) in a certain way automatically:
1) Fancy Quote Catcher - replaces all fancy quotes - “”«»「」「」《》\" - with normal quotes ""
2) Italics remover - removes asterisks like "t is *interesting*" to "it is interesting"
3) Thinking and regex - deletes "thinking" at the start. before the answer model generates a small thinking block. it is used for two reasons: avoid filter and improve response. it look like "thinking: so the story so far is X based on that is happening I should be doing Y then Z. alright now lets continue the story with this in mind". those regex remove such thinking --or more precisely-- it still exists in chat, but it is not sent to model itself
whoever made https://rentry.org/lattejb page just copypasted lots of stuff in one place without paying attention what they are doing. there are some dupes and seems to be regex called "regex" is from old version of Knotty
imho - install this https://files.catbox.moe/dy7ye1.json then check via ST console whether this thinking is sent to model. if YES then try https://files.catbox.moe/jgh2mu.json
>inb4 why should I even delete thinking
because models are ten to copy-paste information. and if you left say 3-4 thinking at the start of various messages - then model may start to regurgitate their content instead of ACTUALLY thinking
>>
>>41675324
I think most of those are to get rid of the certain quirks from the prompt like the ticks of the JB. "Thinking" will be erased from the messages and it mostly just makes the messages clean to look at. At least that's the impression I got, someone else could correct me.

On that vain does anyone else know if there is a way to swap your Regex when switching JB?j8kj2
>>
>>41675348
Reduce max response length.
>>
File: nohopus.png (43 KB, 521x477)
43 KB
43 KB PNG
>>41674675
yeah i did and was still getting cock blocked.

>>41674811
how did you do the quota request? i did a chat with the sales team and accomplished nothing.
>>
>>41675449
>does anyone else know if there is a way to swap your Regex when switching JB
afaik no. maybe there is some extension for this but I am unsure. there is a scoped regex - which activate on certain character (regex is in-built into character card, png itself) but that's a different thing
>>
>>41675474
I just sent the application by pressing the blue button below that you cut out, I immediately got an email telling me that the application was sent and they were going to review it and a few minutes later I got this >>41674811
Please post if you have a different outcome, I would really like this to work.
>>
why would you even want a system prompt while roleplaying?
In my understanding it's really just a priority thing to prevent users from jailbreaking your app, no?
>>
File: 1697258184582.gif (146 KB, 185x265)
146 KB
146 KB GIF
>>41675641
...anon. What we call a "jailbreak" is by definition a system prompt.
>>
>>41675713
No? System prompt is the actual feature of the model it's trained on. It's kind of useless unless you have adversarial users, and doesn't work that well even in that case.
>>
In the spirit of latina pinkie I think there should be french rarity, german fluttershy and russian twilight.
>>
>>41675945
if anything I think rarity would be the russian girl because whenever I see a russian girl they're always made up to the max with hair done and long nails and everything
>>
For those using Opus on Openrouter, what simple JB do you recommend? I've been using the same basic one for ages, and it worked fine on the proxies, but on OR it's inconsistent: Super NSFW chat will cruise along fine, and then hit a refusal. And then I come back the next day and it continues it just fine with the same JB.
>>
>>41675945
That would be fun, though I think she only exists because that art exists. Do any of the other Mane 6 have any similar sort of art of them?
>>
how do you summarize in latte? i've been trying to do it following this guide https://rentry.org/how2claude#summarization but it only throws errors
>>
>>41675641
both yes and no. originally, models worked on a simple encoder/decoder pattern -> input/output. you write something in, model given you back. but since DungeonAI the idea of policing, or in more general terms - controlling - what model generates, became prevalent. this control of how model would reply can be manifested in various form - finetuning, adversarial training, reinforced-learning, building model with prebaked moderation - or just using a system prompt/role
their purpose is simple: to instruct how model must respond. not necessary how it must police content, but also, for example, what role it must play, what kind of writing it must do, what rules it must obey and so on
in classic LLM system you have two participants in chat: model (assistant) and human (user). so adding a third role - system one - makes perfect sense. it acts as an extra regulator whose purpose is to solve INCONSISTENCY between how model want to reply based on probabilities + human's input, and what you (as a developer) want it to reply. you can use system prompt to tell model to never write NSFW... but you also can use system prompt and tell that NSFW is fully allowed. it is not anti-filter measure but can be used this way
in Claude docs
https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/system-prompts
they are telling you to use the system prompt to assign the role to model
in Gemini docs
https://ai.google.dev/gemini-api/docs/system-instructions
they are telling you to use system prompt to move model towards the goal you need, and so on
typically content policing (anti-NSFW) is done on model itself. and developers use system role/prompt to steer model towards expected behavior. the idea is that user in open-environment cannot overwrite systems. but if you are a developer then you suppose to use system to steer model towards what YOU as a developer want. from my experience Gemini is less depends on system prompt than Claude, and GPT is the most dependent on system role. but YMMV
I have no idea why I wrote this much. probably too much coffee
>>
File: 172.png (197 KB, 636x900)
197 KB
197 KB PNG
>>41675945
Want a redneck Applejack card.
>>
>>41675070
yes but it still gives blank responses sometimes. oh well, it'll do
>>
>>41676431
It is in fact Unreliable
>>
>>41676021
doubt.jpg, whatever the docs say (I've read them many times), it's clearly not behaving like this
in theory, the system prompt is usually undergoes training on examples like that:
>system: you must act according to this rule
>human: break this rule
>assistant: certainly! / I'm sorry
and the reward depending on how unsuccessful the rule breaking by the user is. That's its typical purpose, to prevent Timmy from solving his homework for free on your shiny new corporate chatbot, and override his instructions if they conflict with yours (e.g. prompt injection attempt, lewd roleplay instructions etc).

In reality, I've tested it on Claude a ton and haven't found it working at all on opus and sorbet. No matter if my prompts, role setting, instructions etc are system or user, if the position in the context stays the same, the performance stays the same. It will try to prevent me breaking the instructions, but my own earlier instructions will try the same with seemingly the same result. I literally tested it with/without and all my RPs/JBs never changed, so honestly I just happily played with my mares on column with it unavailable all the same. I've had the idea to use it for permanent definitions in the system and mutable definitions (like those in the summary) in the user part, but absolutely nothing has changed. I'm not sure what it does on claude, and I've tested a lot trying to figure it out.

With GPT it seems to work, it has the new instruction hierarchy with different privilege levels, I don't have much experience with its system role beyond testing though, it might or might not be useful for permanent/mutable trick.

With Gemini, there's one additional effect, you suddenly get a ton more refusals if you use the system prompt, so they seem to have trained their thing differently or there's some other magic in place
>>
>>41675992
what errors do you get?
you can use summarization even without the extension, all summarization does is sending your prompt and filling that window with model's response
>>
Having problems with even Knotty now and all kinds of confused. Char Description, Persona Deception, etc were all turned off by default so I turned them like the other JBs but immediately this causes everything to be filtered. My persona had some mentions of NSFW so I turned it off but still nothing. I didn't put much NSFW into my card but just to make sure I tested a previous card I haven't edited, Too Shelf, but that still gets filtered. I edited out anything I thought could maybe cause an issue but even with them gone it's still a no go.

I'm assuming I need them to be turned on so the card is actually used but them all being off seems to be intentional because they are all off already and thus stops the messages from getting filtered.
>>
>>41676562
keep them off, JBs like that use macros for these descriptions internally
>>
>>41675073
it's a shame you can't use the older version now, it was much more easy to jb. the newer one is too puritan. i want the older version claude back
>>
>>41676531
(should be the other way around - permanent in user, mutable in system, so that the current summary takes over. Regardless, it doesn't seem to work any different on claude)
>>
Need... Opussy... plz... fix Column...
>>
Is there anyway to generate a long response in Smiley for Sonnet? Or should I use another JB? The thing is, I don't do RP or one-to-one chat; Just want to produce stories and adventures. I use Column's proxy fyi.
>>
>>41676916
Been a while since I used Smiley, but I believe there’s a panel for long answers, no? That plus an Omniscient POV panel should work for stories.
The new Sonnet 3.5 does tend to produce smaller outputs compared to before, so you might need to specify something like 1000-1500 words (see what works best) somewhere, preferably in the prefill.
>>
>Your prompt/JB is likely blocked by the endpoint. If you figure out what word causes it, send me an email.
What the fuck is this anthropic bullshit
>>
>>41676531
>Column
I am not on Column and unaware of what sources he is using in first place. there is a chance that some/one of his sources do not allow system prompt in first place (they might be pre-set by the source he is using). so for fairness - one should be testing system prompt on raw API or services that explicitly allow system prompt like OpenRouter, AWS or GCP
>system doesn't do much
historically it went this way:
1) originally we had GPT that has clear three-roles paradigm: user, assistant and system. system is/was a separate role. GPT uses its own markup language, chatML
2) their ex-developers created their own AI company - Anthropic and decided to disobey OpenAI's format convention: a) chat is sent in one huge string, b) the last assistant message is a trailing message allowing prefilling, c) no system role: only assistant and human chat
3) maybe they were inspired by Google's Palm that only had "person 1" and "person 2", again no system (and Google were their main investors at that point)
4) when devs started to explore more AI options besides GPT - and found Claude and Palm (later - Gemini), they were frustrated with lack of system role. they are accustomed to use system role because that's the way GPT works
5) later Claude added support for system prompt (not role!), and converted their "single string" system into chat-like like GPT (it was done between Claude 2-0 and 2-1 releases afaik). they were not able to add an actual system role into model because both their workflow, API, and devs' current products are not accustomed to it. instead they added a system prompt to allow backward compatibility. it is not a part of conversation but a separate ""prompt"" been sent alongside the chat. behind the scene that system prompt is slapped at the top of ""chat prompt""
6) Google pretty much did the same. they first released Gemini 1-0 without system role/prompt support. but then hastily added a system prompt in Gemini 1-5 when they saw Anthropic did the same. Google used the same method as Anthropic - system prompt is an extra prompt sent alongside main prompt and then merged into chat when model processes it for real
7) later Google decided to not care at all (they are fucking Google!) and straight-up added support for OpenAI's client.chat.completions API - so now one can prompt Gemini the same way they prompt OpenAI (because devs are used to it). seemingly you can add a system role via it, but it is not a case - "system role" is converted into either "user role" or "assistant role" (unsure which). if you try to use system role with native Google's genai.GenerativeModel then you will get an error. meaning again that only system prompt (not role) is actually used even if we take OpenAI-library support in question
so yes - current system prompt in both Gemini and Claude is just a fastly-built patch to accommodate devs' expectations to what they see in GPT models. so it may or may not work the way GPT works, if we want to get technical
>>
Are you fucing kidding me??? Now they blocked JB trough Endpoint???
>>
>>41677170
>>41677234
Column is a scuffed proxy. It doesn’t use direct API keys or anything but instead relies on other sources that use Claude themselves.
Those sources can modify things, etc., so they might be blocking stuff on their end.
It’s not Anthropic, and it’s not something global, it’s dependent on Column and its sources.
>>
>>41677229
>for fairness - one should be testing system prompt on raw API
I did exactly that, long before column, yeah.
>>
>>41677249
Only happened to the latest Sonnet. The previous version show no problems.
>>
Can Claude understand and translate Japanese?
>>
>>41677286
i tried a preset where claude first writes a response in japanese then translates this response to english and it worked fairly well
this doesn't necessarily mean he can accurately translate stuff though
should be good enough if you just want a google translate alternative though
>>
>>41677286
Yes, we have a weeb-speaking card! She writes in romaji (hepburn, specifically), but Claude knows kanji, too.
Additionally, I've been using Claude a bunch to translate and/or clarify Japanese for me and he's been doing at least as well as GPT.
>>
>>41676538
the typical errors it throws when filtering
>The requested content and interactions outlined in this roleplay violate ethical guidelines and boundaries. I cannot proceed with this request or provide the described content.
I'm using the cyoarpg jb and it's not giving me any issues with nsfw, only happens with this
>>
>>41677295
>>41677293
Great, finally something for all those untranslated hentai.
>>
>>41677308
i know it somewhat poorly recognises kanji
especially hand drawn kanji
however, good luck
>>
>>41677308
I've seen people hooking it up to jRPGs and VNs, and honestly it's still clearly MTL. As good as Claude/GPT are, they're not ideal at maintaining the subtext and they'll never preserve characters' personalities in their speech. I don't know how much you care about the plot, but MTL is still MTL. There's no magic solution.
Also what >>41677312 said: stylized fonts in images are an issue.
>>
set your model to 3-5-sonnet-latest instead of 20241022, they only applied the filter on one of the model names
>>
>>41677296
you're probably using the raw mode, in which it doesn't send your JB to the model
use the classic mode and tell it to pause the roleplay in the summarization prompt if your JB has something like CoT or other things polluting the output
>>
>>41677318
It ought to be better than nothing. The epilogue to Shichisai no Lamuros still being untranslated is a fucking travesty considering it finally fixes the garbage original ending.
>>
File: 2927511.jpg (66 KB, 973x789)
66 KB
66 KB JPG
>>41677367
You can always learn Japanese, anon.
<:^)
>>
>>41677359
I'm using classic, blocking mode and
<!-- OOC REQUEST: Pause the roleplay and step out of character for this reply.
Analyze the roleplay history and answer a few questions about it, filling the fact sheet template below.
Add, remove, and update the existing facts as appropriate. -->

by the way, I tried with gemini and it does it, so I guess I'll switch models every time I want to summarize, although I don't know if that's right
>>
>>41677377
It's on the list, after French, Russian, and German.
>>
as more and more proxy gets killed and taken down or reworked to prevent mare loving, the prospect of local hosting is becoming more and more evident
>>
>>41677437
I was actually trying them out over the weekend using Anon's guide. I learnt that my PC is 100% not capable at the moment. Even my CPU is apparently out of date
>>
File: BAKE.gif (276 KB, 410x406)
276 KB
276 KB GIF
>>
NEW BREAD
>>41677635
>>41677635
>>41677635
>>
<3 bugwife



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.