Perform SFT while keeping only samples with reward 1
synthetic + SFT + reward filtering loop
ReSTEM
Creator
Creator
Seonglae ChoCreated
Created
2025 Oct 14 22:30Editor
Editor
Seonglae ChoEdited
Edited
2025 Oct 14 22:31Refs
Refs
Seonglae ChoSeonglae Cho