ReDiST starts from the target student's own reasoning attempts on evidence-rich QA tasks.
Tailored reasoning traces improve student learning beyond answer-only training
ReDiST revises and selects student-generated traces based on student-specific answer recovery, boosting source and transfer performance on multi-hop QA.
Big Tech
Shaoang Li · Yanhang Shi · Yufei Li · Mingfu Liang · Frank Shyu · Sandeep Pandey · +3 more
Stony Brook University · Meta AI
Research Digest··3 min read
The authors introduce ReDiST, a framework that constructs reasoning traces conditioned on a target student by revising its own correct or failed attempts against provided evidence.
§