mattneary / attention
visualizing attention for LLM users
☆163Updated last year
Related projects ⓘ
Alternatives and complementary repositories for attention
- Extract full next-token probabilities via language model APIs☆229Updated 8 months ago
- Improving Alignment and Robustness with Circuit Breakers☆154Updated last month
- A toolkit for describing model features and intervening on those features to steer behavior.☆99Updated last week
- ☆102Updated last month
- Sparse autoencoders☆342Updated last week
- Evaluating LLMs with fewer examples☆134Updated 7 months ago
- Code accompanying "How I learned to start worrying about prompt formatting".☆95Updated last month
- Function Vectors in Large Language Models (ICLR 2024)☆119Updated last month
- Code for In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steering☆144Updated last month
- Create feature-centric and prompt-centric visualizations for sparse autoencoders (like those from Anthropic's published research).☆157Updated last month
- Steering vectors for transformer language models in Pytorch / Huggingface☆65Updated last month
- ☆107Updated this week
- Steering Llama 2 with Contrastive Activation Addition☆97Updated 5 months ago
- Erasing concepts from neural representations with provable guarantees☆209Updated last week
- Controlled Text Generation via Language Model Arithmetic☆212Updated 2 months ago
- Code and results accompanying the paper "Refusal in Language Models Is Mediated by a Single Direction".☆123Updated last month
- Training Sparse Autoencoders on Language Models☆469Updated this week
- MiniCheck: Efficient Fact-Checking of LLMs on Grounding Documents [EMNLP 2024]☆103Updated last month
- Scripts for generating synthetic finetuning data for reducing sycophancy.☆107Updated last year
- code for training & evaluating Contextual Document Embedding models☆117Updated this week
- RuLES: a benchmark for evaluating rule-following in language models☆211Updated last month
- ModuleFormer is a MoE-based architecture that includes two different types of experts: stick-breaking attention heads and feedforward exp…☆216Updated 7 months ago
- LLM experiments done during SERI MATS - focusing on activation steering / interpreting activation spaces☆78Updated last year
- ☆199Updated this week
- ☆101Updated 3 months ago
- Code repository for the c-BTM paper☆105Updated last year
- Scaling Data-Constrained Language Models☆321Updated last month
- Functional Benchmarks and the Reasoning Gap☆78Updated last month
- Fast & more realistic evaluation of chat language models. Includes leaderboard.☆183Updated 10 months ago
- [EMNLP 2023] Adapting Language Models to Compress Long Contexts☆277Updated 2 months ago