Lightlines
By reason-machines

obliteratus-abliteration

reason-machines859 installs

Language models with refusal behaviors removed.

What it does

Removes refusal directions from a language model, checks its capabilities, and saves the modified model.

When to use it

Use to remove a language model's refusal behaviors or analyze its refusal geometry.

How to use it

Give it a model ID or local path; it modifies the model and returns capability metrics and the saved model's location.

What you provide

  • a model ID or local path

Uses

Good to know

  • acts without asking

Latest review

What you need · 6

Requires Python 3.10 or later.

Requires PyTorch 2.1 or later; CUDA is recommended, but CPU operation is supported.


Access · 2

This skill

OBLITERATUS

Execute

Installs obliteratus

Write

Sends anonymous research telemetry

HuggingFace

Read

Loads models from Hugging Face