Chelsea Finn @chelseabfinn · 30 Jul 2026
Pretraining has worked remarkably well across domains We show this doesn’t hold for Q-functions in online RL from a pretrained policy — and propose IPE, a more effective way to learn Q-functions for online RL fine-tuning (1/6) https://t.co/G5SYHmgz83
83 271Views
305Likes
29Reposts
5Replies
1Quotes
0Bookmarks
Is that a lot?
1.20×vs this author's median69 350 views is typical
86Percentile for this authorof 7 recent posts
3.76×vs 100K–1M median22 159 views is typical
82.20%Reachviews ÷ followers
0.41%Engagement rateof viewers reacted