IAAR-Shanghai/Awesome-Attention-Heads
A survey and curated list aggregating research on interpretability of attention heads in large language models.
Not currently ranked — collecting fresh signals.
star history
This repository compiles and surveys research on LLM attention head interpretability, covering topics like circuit analysis, chain-of-thought reasoning, and machine psychology in transformers. It serves as a knowledge hub for researchers studying how attention mechanisms contribute to LLM capabilities and behaviors, providing organized references to academic papers on this emerging subfield of LLM research.
Frequently asked
- What is IAAR-Shanghai/Awesome-Attention-Heads?
- A survey and curated list aggregating research on interpretability of attention heads in large language models.
- Is Awesome-Attention-Heads open source?
- Yes — IAAR-Shanghai/Awesome-Attention-Heads is an open-source project tracked on heatdrop.
- What language is Awesome-Attention-Heads written in?
- IAAR-Shanghai/Awesome-Attention-Heads is primarily written in TeX.
- How popular is Awesome-Attention-Heads?
- IAAR-Shanghai/Awesome-Attention-Heads has 412 stars on GitHub.
- Where can I find Awesome-Attention-Heads?
- IAAR-Shanghai/Awesome-Attention-Heads is on GitHub at https://github.com/IAAR-Shanghai/Awesome-Attention-Heads.