frankaging
update
558908a

A newer version of the Gradio SDK is available: 5.16.0

Upgrade
metadata
title: SDL-ReFT-cr1
emoji: 🫠
colorFrom: red
colorTo: indigo
sdk: gradio
sdk_version: 5.13.1
app_file: app.py
pinned: false
suggested_hardware: a10g-small

Model conditioned steering with supervised dictionary learning (SDL).

This is a demo of model steering with Supervised Dictionary Learning (SDL) using AxBench-ReFT-r1-16K which hosts steering vectors for 16K concepts.