RewardHacking

AI & ML interests

None defined yet.

Recent Activity

tongliuphysics  authored a paper 22 days ago

Temperature-scaling surprisal estimates improve fit to human reading times -- but does it do so for the "right reasons"?

tongliuphysics  authored a paper 22 days ago

FocalPO: Enhancing Preference Optimizing by Focusing on Correct Preference Rankings

tongliuphysics  authored a paper about 1 year ago

Multimodal Pragmatic Jailbreak on Text-to-image Models

View all activity

Team members 2