nrimsky / LM-exp
LLM experiments done during SERI MATS - focusing on activation steering / interpreting activation spaces
☆91Updated last year
Alternatives and similar repositories for LM-exp:
Users that are interested in LM-exp are comparing it to the libraries listed below
- Steering Llama 2 with Contrastive Activation Addition☆125Updated 9 months ago
- For OpenMOSS Mechanistic Interpretability Team's Sparse Autoencoder (SAE) research.☆95Updated this week
- A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity.☆69Updated this week
- Algebraic value editing in pretrained language models☆63Updated last year
- ☆80Updated 7 months ago
- Steering vectors for transformer language models in Pytorch / Huggingface☆90Updated 2 weeks ago
- Röttger et al. (NAACL 2024): "XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Models"☆85Updated 2 weeks ago
- Improving Alignment and Robustness with Circuit Breakers☆189Updated 5 months ago
- datasets from the paper "Towards Understanding Sycophancy in Language Models"☆72Updated last year
- ☆148Updated this week
- Using sparse coding to find distributed representations used by neural networks.☆218Updated last year
- A resource repository for representation engineering in large language models☆109Updated 3 months ago
- General-purpose activation steering library☆49Updated 2 months ago
- ☆192Updated last year
- WMDP is a LLM proxy benchmark for hazardous knowledge in bio, cyber, and chemical security. We also release code for RMU, an unlearning m…☆105Updated 10 months ago
- Repo accompanying our paper "Do Llamas Work in English? On the Latent Language of Multilingual Transformers".☆70Updated last year
- ☆210Updated 5 months ago
- ☆89Updated last year
- Sparse probing paper full code.☆55Updated last year
- AI Logging for Interpretability and Explainability🔬☆105Updated 9 months ago
- ☆123Updated 3 months ago
- Function Vectors in Large Language Models (ICLR 2024)☆142Updated 5 months ago
- ☆159Updated this week
- ☆38Updated last year
- ☆112Updated 7 months ago
- A library for efficient patching and automatic circuit discovery.☆56Updated 3 weeks ago
- ☆21Updated 2 weeks ago
- [NAACL'25 Oral] Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering☆50Updated 3 months ago
- Code release for "Debating with More Persuasive LLMs Leads to More Truthful Answers"☆101Updated 11 months ago
- [ICML 2024] Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications☆71Updated 2 weeks ago