This website requires javascript to properly function. Consider activating javascript to get access to all site functionality.
AI ALIGNMENT FORUM
AF
Login
Wes Gurnee
OR PhD student at MIT working on interpretability.
Find out more here: https://wesg.me/
Posts
Sorted by New
77
Refusal in LLMs is mediated by a single direction
8mo
43
52
SAE reconstruction errors are (empirically) pathological
9mo
1
19
Finding Neurons in a Haystack: Case Studies with Sparse Probing
2y
1
Wiki Contributions
Comments
Sorted by
Newest