microsoft / mechanistic-error-probeLinks
A mechanistic approach for understanding and detecting factual errors of large language models.
☆49Updated last year
Alternatives and similar repositories for mechanistic-error-probe
Users that are interested in mechanistic-error-probe are comparing it to the libraries listed below
Sorting:
- Code for the ICLR 2024 paper "How to catch an AI liar: Lie detection in black-box LLMs by asking unrelated questions"☆71Updated last year
- Steering vectors for transformer language models in Pytorch / Huggingface☆140Updated 11 months ago
- Investigating the generalization behavior of LM probes trained to predict truth labels: (1) from one annotator to another, and (2) from e…☆28Updated last year
- ☆112Updated 11 months ago
- datasets from the paper "Towards Understanding Sycophancy in Language Models"☆102Updated 2 years ago
- Code for reproducing our paper "Not All Language Model Features Are Linear"☆83Updated last year
- ☆82Updated last year
- Code release for "Debating with More Persuasive LLMs Leads to More Truthful Answers"☆124Updated last year
- A library for efficient patching and automatic circuit discovery.☆88Updated last month
- Sparse and discrete interpretability tool for neural networks☆64Updated last year
- We develop benchmarks and analysis tools to evaluate the causal reasoning abilities of LLMs.☆137Updated last year
- ☆99Updated last year
- Repository for the paper Stream of Search: Learning to Search in Language☆153Updated last year
- A repository for transformer critique learning and generation☆89Updated 2 years ago
- ☆57Updated 2 years ago
- Functional Benchmarks and the Reasoning Gap☆89Updated last year
- Open source replication of Anthropic's Crosscoders for Model Diffing☆63Updated last year
- Repo for: When to Make Exceptions: Exploring Language Models as Accounts of Human Moral Judgment☆38Updated 2 years ago
- Code for In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steering☆198Updated 11 months ago
- Notebooks accompanying Anthropic's "Toy Models of Superposition" paper☆135Updated 3 years ago
- Code and Data Repo for the CoNLL Paper -- Future Lens: Anticipating Subsequent Tokens from a Single Hidden State☆20Updated 3 months ago
- Stanford NLP Python library for benchmarking the utility of LLM interpretability methods☆163Updated 7 months ago
- Evaluation of neuro-symbolic engines☆41Updated last year
- ☆138Updated last year
- ☆68Updated last year
- ☆152Updated 5 months ago
- PyTorch library for Active Fine-Tuning☆96Updated 4 months ago
- Erasing concepts from neural representations with provable guarantees☆243Updated last year
- ☆105Updated last year
- PASTA: Post-hoc Attention Steering for LLMs☆134Updated last year