Orion-zhen/abliteration
Make abliterated models with transformers, easy and fast
This tool helps developers and researchers modify Large Language Models (LLMs) to reduce their tendency to refuse user inputs, especially for sensitive topics. You provide a pre-trained transformer-based LLM, along with 'harmful' and 'harmless' prompt examples. The output is a modified version of your LLM that is less likely to explicitly refuse queries, while aiming to preserve other capabilities. This is for AI practitioners working with custom or open-source LLMs.
128 stars.
Use this if you need to fine-tune an LLM to be less overtly restrictive or refusing in its responses, without completely uncensoring it, and you understand the implications of model modification.
Not ideal if you are looking for a complete uncensoring solution for an LLM or if you don't have the technical expertise to work with transformer models directly.
Stars
128
Forks
54
Language
Python
License
GPL-3.0
Category
Last pushed
Dec 09, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/Orion-zhen/abliteration"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related models
ModelTC/LightCompress
[EMNLP 2024 & AAAI 2026] A powerful toolkit for compressing large models including LLMs, VLMs,...
p-e-w/heretic
Fully automatic censorship removal for language models
YerbaPage/LongCodeZip
LongCodeZip: Compress Long Context for Code Language Models [ASE2025]
locuslab/wanda
A simple and effective LLM pruning approach.
tommasomncttn/mergenetic
Flexible library for merging large language models (LLMs) via evolutionary optimization (ACL 2025 Demo).