Sidharth Baskaran comments on Fluent dreaming for language models (AI interpretability method)

Sidharth Baskaran 16 Nov 2024 23:42 UTC
1 point
0
Cool followup work here!
https://www.lesswrong.com/posts/hMBTaFvAzdMNnj29c/evolutionary-prompt-optimization-for-sae-feature