Saisab Sadhu's picture

Building on HF

1

Saisab Sadhu

saisabs

·

https://saisabsadhu.github.io/

AI & ML interests

Mechanistic Interpretation, Unlearning

Recent Activity

updated a model 6 minutes ago

Lexsi/audit-steer-RefusalDirectionModel-llama31-8b-dolly

published a model 6 minutes ago

Lexsi/audit-steer-RefusalDirectionModel-llama31-8b-dolly

updated a model 21 minutes ago

Lexsi/audit-steer-CAAModel-llama31-8b-dolly

View all activity

Organizations

upvoted a paper about 17 hours ago

Forgetting That Sticks: Quantization-Permanent Unlearning via Circuit Attribution

Paper • 2605.15138 • Published 9 days ago • 6