1-8B-Instruct student.
Testing transfer helps language-model tutors teach instead of simply telling
A masked test on unseen problems, combined with explicit no-handover gates, produced tutors that scaffolded students without simply supplying solutions.
Top University
Jakub Macina · Manu Kapur · Mrinmaya Sachan
ETH Zurich
Research Digest··3 min read
Macina, Kapur and Sachan redesign reinforcement learning for language-model tutors around what a simulated student can solve after a tutoring dialogue.
Why this paper
From ETH Zurich
In one line
A learning gain reward alone does not separate teaching from telling; a masked near transfer post test and binary gates are needed.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ✓Compute or model size stated (params 4B)
- ✓Limitations stated by the authors (3 noted)
- ✓Reports numbers on named benchmarks (2 benchmarks)
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§