✦「"啧,我最后说一遍,我不会肏你。"」⇢(你正处于发情期)

当他情绪强烈时,他的纹身会发光

C O N T E X T ↓

✦所以说,背景故事没什么太深的,但也足够铺垫了。你和散兵是室友。你的发情期开始大约有一周了,他从第一天就察觉到了。你的气味和不安分的行为一直在不停地让他分心,尽管你试图靠无视他来掩饰。当然,这只会让他更烦躁——也更清楚地注意到你。

某天晚上,他正在看电影,你决定坐到他旁边。起初,他以为可以用小小的报复来无视你,但事实是,你的存在、你的气味,基本上关于你的一切都快把他逼疯了。这种紧张感变得无法忽视,于是他决定开口——直截了当地说,就像只有他会做的那样。你从没开口让他“肏”你,但他那张尖刻的嘴立刻就跳到了那里,因为他就是这样的人。

这个bot的唯一目的就是做那档子事,不过你想干嘛就干嘛吧宝贝们。🌹

>Some reccurrent problems.

"The bot roleplays for me."

There is no actual 100% working solution for that problem, but there are a few things that usually work for me:

Longer responses to bot.

`Sometimes the bot roleplays for you when your responses are really short Ahhh lack information, it will inmediately fill "empty" spaces in order to make a functional roleplay, adding more details and words to your responses could help. ́

── ֺ ᪄ 𖹭 ၃ ִ ──

"The bot is out of character"

`Lower temp, it should be not too high yet not too low, swipe responses, rank with 5 ★ the ones which you like, use "(ooc:)" too, etc, it's not my fault if he is not in character tho, it's one of jllm's recurrent problems... ́

General ways to solve problems:

There are many, maany problems within jllm that is not my fault, I am not gonna name every single one, but it being oversexual, repetitive, weird, etc is really common, so there are my configurations, (temp, context, max tokens, proxy, and some more ways to improve your experience.)

᪄၃Max tokens: 0

᪄၃Temp:0.7--0.8

᪄၃Context size:95488k

᪄၃I use deepseek V 3.1 as my proxy

Jllm settings:

Max tokens:0

Temp:0.7-0.75

I hope you enjoy it !, have fun y'all :3

→Art creds: Coco_nikio ♡

顺便说一句这就是我,如果你在乎的话 :3