IBM / sae-steeringLinks
Code to enable layer-level steering in LLMs using sparse auto encoders
☆19Updated last month
Alternatives and similar repositories for sae-steering
Users that are interested in sae-steering are comparing it to the libraries listed below
Sorting:
- Code for reproducing our paper "Not All Language Model Features Are Linear"☆75Updated 7 months ago
- [NeurIPS 2024 Spotlight] Code and data for the paper "Finding Transformer Circuits with Edge Pruning".☆56Updated 3 months ago
- Code for the ICLR 2024 paper "How to catch an AI liar: Lie detection in black-box LLMs by asking unrelated questions"☆70Updated last year
- Stanford NLP Python library for benchmarking the utility of LLM interpretability methods☆95Updated 3 weeks ago
- Improving Steering Vectors by Targeting Sparse Autoencoder Features☆21Updated 7 months ago
- PaCE: Parsimonious Concept Engineering for Large Language Models (NeurIPS 2024)☆37Updated 7 months ago
- Investigating the generalization behavior of LM probes trained to predict truth labels: (1) from one annotator to another, and (2) from e…☆27Updated last year
- Sparse Autoencoder Training Library☆52Updated last month
- ☆65Updated last year
- This repository contains the code and data for the paper "SelfIE: Self-Interpretation of Large Language Model Embeddings" by Haozhe Chen,…☆50Updated 6 months ago
- Algebraic value editing in pretrained language models☆65Updated last year
- ☆85Updated 10 months ago
- ☆14Updated last year
- Open source replication of Anthropic's Crosscoders for Model Diffing☆55Updated 8 months ago
- ☆37Updated last month
- ☆23Updated 5 months ago
- A library for efficient patching and automatic circuit discovery.☆67Updated 2 months ago
- ☆44Updated last year
- Code release for "Debating with More Persuasive LLMs Leads to More Truthful Answers"☆109Updated last year
- [ICML 2025] Flow of Reasoning: Training LLMs for Divergent Reasoning with Minimal Examples☆95Updated 2 weeks ago
- Steering vectors for transformer language models in Pytorch / Huggingface☆110Updated 4 months ago
- ☆34Updated 5 months ago
- Multi-Layer Sparse Autoencoders (ICLR 2025)☆22Updated 4 months ago
- Steering Llama 2 with Contrastive Activation Addition☆159Updated last year
- Code for reproducing our paper "Low Rank Adapting Models for Sparse Autoencoder Features"☆10Updated 2 months ago
- General-purpose activation steering library☆81Updated last month
- Function Vectors in Large Language Models (ICLR 2024)☆170Updated 2 months ago
- ConceptVectors Benchmark and Code for the paper "Intrinsic Evaluation of Unlearning Using Parametric Knowledge Traces"☆35Updated 4 months ago
- ☆31Updated last year
- LLM experiments done during SERI MATS - focusing on activation steering / interpreting activation spaces☆94Updated last year