Hacker News new | past | comments | ask | show | jobs | submit | from login

Show HN: Llama 3.2 Interpretability with Sparse Autoencoders (github.com/paulpauls) 442 points by PaulPauls 16 hours ago | past | 58 comments

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact Search: