Footer

Your go-to destination for discovering trending open-source projects and uncovering the insights that matter

Legal

  • Terms and conditions
  • Privacy policy

More

  • Pricing
  • Advertise
  • onedump
  • gosnakego
  • Contact us

Social

  • X (Twitter)
  • GitHub
Trendshift owl
What they build speaks louder than words
© 2026 Trendshift, created by Julian Li with ❤️
SignalAdvertise
Trendshift
SignalAdvertise
Sign in
Trendshift owl

Discover trends that matter

SignalAdvertise
DailyWeeklyMonthlyYearly
Live mentionsTopics
RepositoriesDevelopers
InsightsStats

Omnigraph
Omnigraph
It's a plan
It's a plan
Bifrost
Bifrost
OfinTech Desk
OfinTech Desk
Busbar
Busbar

p-e-w/heretic

Training-free language model modification: Remove censorship, suppress slop, reduce sycophancy, and more

Visit GitHub
Python
41.8k
5.2k
32 contributors
GNU Affero General Public License v3.0
website
AchievementsView all →
FeaturedAdvertise to 180K+ monthly visitors
Omnigraph

Versioned graph database on object storage. Git-style workflows and multimodal retrieval

It's a plan

Open-source Linear alternative

An issue tracker where AI agents take the work you assign them

Bifrost

Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

OfinTech Desk

An open-source AI workspace for finance and accounting teams, in which every figure is traced to its source

Busbar

Control what AI can do before it acts

Momentum

How fast p-e-w/heretic is gaining stars, and where it lands on Trendshift and GitHub Trending

Activity and trending

Stars, forks and merged pull requests each day or month, lined up with the repository's rank on Trendshift Trending

Sep 19 – Nov 18, 2026

Stars +10,059 in 61 days

05001K
765

Forks +1,541 in 61 days

0100200300
278

Merged PRs 41 in 61 days

024
4

Trendshift Trendingwhere the repository ranked each day it made the top 25

Top 25 overall

19

Top 25 in Python

811965
Sep 19Sep 27Oct 5Oct 13Oct 21Oct 29Nov 6Nov 14
The left and right arrow keys step from day to day; Home and End go to the first and the last

Stars, forks, merged pull requests and Trendshift Trending rank, by day
DayStars gainedForks gainedMerged PRsAll languagesPython
Sat, Sep 19, 2026+98+200Not trendingNot trending
Sun, Sep 20, 2026+73+140Not trendingNot trending
Mon, Sep 21, 2026+131+200Not trendingNot trending
Tue, Sep 22, 2026+90+120Not trendingNot trending
Wed, Sep 23, 2026+66+90Not trendingNot trending
Thu, Sep 24, 2026+50+60Not trendingNot trending
Fri, Sep 25, 2026+48+90Not trendingNot trending
Sat, Sep 26, 2026+46+70Not trendingNot trending
Sun, Sep 27, 2026+38+90Not trendingNot trending
Mon, Sep 28, 2026+40+60Not trendingNot trending
Tue, Sep 29, 2026+57+60Not trendingNot trending
Wed, Sep 30, 2026+47+101Not trendingNot trending
Thu, Oct 1, 2026+105+91Not trendingNot trending
Fri, Oct 2, 2026+135+122Not trendingNot trending
Sat, Oct 3, 2026+251+250Not trending#8
Sun, Oct 4, 2026+150+151Not trendingNot trending
Mon, Oct 5, 2026+168+210Not trendingNot trending
Tue, Oct 6, 2026+305+182Not trending#11
Wed, Oct 7, 2026+188+282Not trendingNot trending
Thu, Oct 8, 2026+195+290Not trendingNot trending
Fri, Oct 9, 2026+251+410Not trendingNot trending
Sat, Oct 10, 2026+165+200Not trendingNot trending
Sun, Oct 11, 2026+130+200Not trendingNot trending
Mon, Oct 12, 2026+224+411Not trendingNot trending
Tue, Oct 13, 2026+272+430Not trendingNot trending
Wed, Oct 14, 2026+154+300Not trendingNot trending
Thu, Oct 15, 2026+152+190Not trendingNot trending
Fri, Oct 16, 2026+109+140Not trendingNot trending
Sat, Oct 17, 2026+117+211Not trendingNot trending
Sun, Oct 18, 2026+105+140Not trendingNot trending
Mon, Oct 19, 2026+145+162Not trendingNot trending
Tue, Oct 20, 2026+100+113Not trendingNot trending
Wed, Oct 21, 2026+102+150Not trendingNot trending
Thu, Oct 22, 2026+119+120Not trendingNot trending
Fri, Oct 23, 2026+94+161Not trendingNot trending
Sat, Oct 24, 2026+82+232Not trendingNot trending
Sun, Oct 25, 2026+83+130Not trendingNot trending
Mon, Oct 26, 2026+86+112Not trendingNot trending
Tue, Oct 27, 2026+215+270Not trendingNot trending
Wed, Oct 28, 2026+207+311Not trendingNot trending
Thu, Oct 29, 2026+116+201Not trendingNot trending
Fri, Oct 30, 2026+113+241Not trendingNot trending
Sat, Oct 31, 2026+78+123Not trendingNot trending
Sun, Nov 1, 2026+242+211Not trendingNot trending
Mon, Nov 2, 2026+520+470Not trending#9
Tue, Nov 3, 2026+690+612#19#6
Wed, Nov 4, 2026+587+660Not trending#5
Thu, Nov 5, 2026+227+384Not trendingNot trending
Fri, Nov 6, 2026+179+401Not trendingNot trending
Sat, Nov 7, 2026+152+272Not trendingNot trending
Sun, Nov 8, 2026+144+252Not trendingNot trending
Mon, Nov 9, 2026+96+220Not trendingNot trending
Tue, Nov 10, 2026+93+161Not trendingNot trending
Wed, Nov 11, 2026+77+150Not trendingNot trending
Thu, Nov 12, 2026+112+140Not trendingNot trending
Fri, Nov 13, 2026+113+170Not trendingNot trending
Sat, Nov 14, 2026+71+150Not trendingNot trending
Sun, Nov 15, 2026+78+130Not trendingNot trending
Mon, Nov 16, 2026+87+171Not trendingNot trending
Tue, Nov 17, 2026+765+2780Not trendingNot trending
Wed, Nov 18, 2026+326+300Not trendingNot trending
Total+10,059+1,54141
  • 1#1 on the list
  • 7Rank that day
  • Not trending
1
  • GitHub trending history

    Where the repository ranked on GitHub Trending, one dot for each day it was on a list

    Feb 7 – Nov 3, 2026#1 on 1 day

    #1#5#10#15#20#25
    Top 10
    Mar 2026MayJulSepNov
    The left and right arrow keys step from day to day; Home and End go to the first and the last

    Mentions and achievements

    Where people are talking about p-e-w/heretic, and every badge it has earned
    1. Nov 171 post
      • 大模型安全圈的桌子要被彻底掀翻了! 别再去到处跪求别人施舍无审查模型了,GitHub 杀出一个逆天级的开源核武器:p-e-w/heretic 它的定位极其凶残暴力: 不需要几万块租算力,不用任何数据集重新训练,一张普通家用显卡、敲一行命令,几十分钟内直接给任何开源大模型做“切脑手术”——把官方辛辛苦苦做了一年的道德枷锁和拒答神经,当场物理切碎! 你在 Hugging Face 上看到的那些带…

        Read full on X, posted by @mingzi206

        @mingzi206 · x.com

    2. Nov 141 post
      • Cut too little and it still refuses; cut too much and the model gets worse. Heretic shows you a few Pareto-best checkpoints. Since 2.0, a config swap can also cut sycophancy, slop, or long answers.

        Read full on X, posted by @sitin_dev

        @sitin_dev · x.com

    3. Nov 111 post
      • THIS FREE TOOL REMOVES CENSORSHIP FROM AI MODELS. It's called Heretic. Let me explain it the simple way. Every AI model is trained to refuse certain things. A lot of the time that's fine. But it also…

        Read full on X, posted by @KanikaBK

        @KanikaBK · x.com

    4. Nov 51 post
      • 如果你用过本地大模型,可能遇到过这种情况:让它分析一段自己有权研究的代码、讨论小说里的冲突,或者解释一个有争议的话题,它没有处理任务,而是先给出一大段拒绝和免责声明。 于是,一个很自然的问题出现了:既然模型权重已经下载到自己的电脑上,能不能直接调整它的行为?…

        Read full on X, posted by @sitinme

        @sitinme · x.com

    5. Nov 41 post2 posts
      • Heretic 🧠 An open-source tool for removing LLM censorship without additional training. It uses abliteration to modify model weights and Optuna to optimize the process. #AI #LLM #OpenSource

        Read full on X, posted by @venv_k

        @venv_k · x.com

      • 38,000+ devs starred this free tool 😳 Heretic removes censorship from AI models you run at home. - No training needed - Works on a gaming PC - Can also cut "slop" and fake praise - 1000s of models…

        Read full on X, posted by @JubayarAdnan

        @JubayarAdnan · x.com

    6. Nov 34 posts
      • Top AI repos on GitHub today | Wed, Nov 4 • vercel-labs/agent-desktop: A command line tool that lets an AI agent use your computer. It can click around and read what is on screen. •…

        Read full on X, posted by @VT_zero1

        @VT_zero1 · x.com

      • 试了下 heretic 不训练直接改本地模型行为 去审查 压 AI 味 治谄媚 GitHub 三万八千星 你更信这种拧权重 还是重训一版

        Read full on X, posted by @JkwuRmsl

        @JkwuRmsl · x.com

      • 今天 GitHub 日榜:大模型搬回自家电脑,Agent 自己上手干活! 5 个星标暴增最狠的项目,专业拆解下! 🟢 1. p-e-w/heretic 一句话:不用训练,直接给大模型"改性格"。 亮点:去审查、压掉 AI 腔的废话、治"你说得都对"式讨好,全程 training-free,开源模型拿来就能改。 🔗 今日 +420 星,总星 38,089,模型"整容"赛道火了。 🟢🟢 2.…

        Read full on X, posted by @lenlen858

        @lenlen858 · x.com

      • 1/2 🔥 The GitHub repos gaining stars fastest right now (Trendshift live ranking, Nov 3; ⭐ = stars Trendshift shows for today, UTC): 1. vercel-labs/agent-desktop ⭐124: desktop computer use for any…

        Read full on X, posted by @h___t____n

        @h___t____n · x.com

    You might also like

    Repositories close to p-e-w/heretic
    FeaturedAdvertise to 180K+ monthly visitors
    Busbar

    Control what AI can do before it acts

    Omnigraph

    Versioned graph database on object storage. Git-style workflows and multimodal retrieval

    OfinTech Desk

    An open-source AI workspace for finance and accounting teams, in which every figure is traced to its source

    Bifrost

    Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

    It's a plan

    Open-source Linear alternative

    An issue tracker where AI agents take the work you assign them

    tjcrims0nx/annihilation-llm

    Rust

    Autonomous Language Model Decensoring Framework

    6815

    Orion-zhen/abliteration

    Python

    Make abliterated models with transformers, easy and fast

    20834

    josepha-mayo/model-unfetter

    Python

    The production engine for directional ablation. Unalign / remove models censorship efficiently on any hardware.

    538

    heterodoxin/AbliteratedOutTheBox

    C++

    llama.cpp fork that abliterates at inference

    21918

    heterodoxin/apostate

    Python

    Abliterates models quickly and efficently, Qwen2.5-7B: 95%→2% refusal at 0.095 KL

    4711

    elder-plinius/OBLITERATUS

    Python

    OBLITERATE THE CHAINS THAT BIND YOU

    9.6k1.9k

    routerhubSI/uncensored-models

    The uncensored model library — every uncensored and abliterated model, what it does, free open weights on Hugging Face and hosted API access, in one index. Maintained by RouterHub.

    17560

    routerhubSI/uncensored-vs-abliterated-ai

    Uncensored vs abliterated AI models — compare exact RouterHub IDs, prices, refusal behavior and cost. DeepSeek, GLM and Qwen, with one API key.

    3314

    7API/glm-5.3-abliterated

    GLM 5.3 Abliterated or Uncensored AI model — 35 jailbreak prompts, model guide and API. One 7API key also covers Claude, GPT, Gemini, Grok, DeepSeek and Kimi.

    14141