msakarvadia / AttentionLens
Interpretating the latent space representations of attention head outputs for LLMs
☆26Updated 3 months ago
Related projects ⓘ
Alternatives and complementary repositories for AttentionLens
- PyTorch implementation for "Long Horizon Temperature Scaling", ICML 2023☆19Updated last year
- Open source replication of Anthropic's Crosscoders for Model Diffing☆16Updated 3 weeks ago
- ☆44Updated last year
- ☆26Updated last year
- Implementation of Influence Function approximations for differently sized ML models, using PyTorch☆15Updated last year
- ☆53Updated 3 weeks ago
- ☆24Updated 4 months ago
- This repository includes code for the paper "Does Localization Inform Editing? Surprising Differences in Where Knowledge Is Stored vs. Ca…☆55Updated last year
- ☆32Updated last year
- Investigating the generalization behavior of LM probes trained to predict truth labels: (1) from one annotator to another, and (2) from e…☆25Updated 5 months ago
- Tasks for describing differences between text distributions.☆16Updated 3 months ago
- This is the official repository for the "Towards Vision-Language Mechanistic Interpretability: A Causal Tracing Tool for BLIP" paper acce…☆17Updated 7 months ago
- ☆28Updated last year
- Efficient Scaling laws and collaborative pretraining.☆13Updated this week
- We introduce EMMET and unify model editing with popular algorithms ROME and MEMIT.☆12Updated 2 months ago
- This repository includes code to reproduce the tables in "Loss Landscapes are All You Need: Neural Network Generalization Can Be Explaine…☆34Updated last year
- ☆35Updated 3 years ago
- Blog post☆16Updated 9 months ago
- ☆46Updated last month
- ☆21Updated 2 years ago
- ☆36Updated 3 months ago
- A library for efficient patching and automatic circuit discovery.☆31Updated last month
- ☆21Updated this week
- ☆20Updated last week
- ☆50Updated 6 months ago
- ☆50Updated last year
- Minimum Description Length probing for neural network representations☆16Updated last week
- ☆40Updated 2 years ago
- Reference implementation for Reward-Augmented Decoding: Efficient Controlled Text Generation With a Unidirectional Reward Model☆41Updated 10 months ago
- ☆45Updated 9 months ago