mitvis / saliency-cards
Saliency Cards are transparency documentation for saliency methods. Learn about new saliency methods or document your own!
☆15Updated last year
Related projects: ⓘ
- The Official Repository for "Bring Your Own Data! Self-Supervised Evaluation for Large Language Models"☆109Updated 11 months ago
- 🧠Starter templates for doing interpretability research☆59Updated last year
- The GitHub repo for Goal Driven Discovery of Distributional Differences via Language Descriptions☆68Updated last year
- NeuroSurgeon is a package that enables researchers to uncover and manipulate subnetworks within models in Huggingface Transformers☆33Updated last month
- Steering Llama 2 with Contrastive Activation Addition☆83Updated 3 months ago
- ☆54Updated last week
- ☆99Updated 10 months ago
- A mechanistic approach for understanding and detecting factual errors of large language models.☆38Updated 2 months ago
- ☆57Updated 5 months ago
- Mechanistic Interpretability Visualizations using React☆175Updated 2 months ago
- ☆75Updated this week
- Erasing concepts from neural representations with provable guarantees☆202Updated 3 months ago
- ☆29Updated 10 months ago
- ☆66Updated last month
- Code for Language-Interfaced FineTuning for Non-Language Machine Learning Tasks.☆120Updated 5 months ago
- Keeping language models honest by directly eliciting knowledge encoded in their activations.☆182Updated this week
- ☆91Updated last month
- We develop benchmarks and analysis tools to evaluate the causal reasoning abilities of LLMs.☆87Updated 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"☆56Updated 3 months ago
- Universal Neurons in GPT2 Language Models☆25Updated 3 months ago
- ☆68Updated 7 months ago
- datasets from the paper "Towards Understanding Sycophancy in Language Models"☆59Updated 10 months ago
- we got you bro☆32Updated last month
- Scaling is a distributed training library and installable dependency designed to scale up neural networks, with a dedicated module for tr…☆38Updated 3 weeks ago
- Steering vectors for transformer language models in Pytorch / Huggingface☆52Updated 2 months ago
- ☆91Updated 5 months ago
- ☆68Updated last month
- A library for efficient patching and automatic circuit discovery.☆18Updated 3 weeks ago
- Using sparse coding to find distributed representations used by neural networks.☆162Updated 10 months ago
- WMDP is a LLM proxy benchmark for hazardous knowledge in bio, cyber, and chemical security. We also release code for RMU, an unlearning m…☆72Updated 4 months ago