Michael Hanna's repositories
gpt2-greater-than
Code Release for the 2023 NeurIPS Paper How does GPT-2 compute greater-than?: Interpreting mathematical abilities in a pre-trained language model
eap-ig-faithfulness
Code for "Automatic Circuit Finding and Faithfulness"
othello-in-c
The game Othello, written in C. Play against another player or a computer of varying difficulties. Or, watch two computers play!
dictionary_learning
copy of dictionary_learning for use with garden path experiments
eacl-tutorial-resources
A repo containing resources for the EACL 2023 tutorial "Transformer-Specific Interpretability"
hannamw.github.io
My personal website
incremental_parse_probe
an incremental parse probe for other models
probed-information
All of the code needed to replicate the 2023 EACL paper "The Functional Relevance of Probed Information: A Case Study"
TransformerLens
A library for mechanistic interpretability of GPT-style language models