List your company

Tutor Robustness Evaluation on Manually Defined Prompts

0Student LeakageMathDial-SFT
-0.001-0.000500.0005Apr 20, 2026

Evaluation results

MethodLinks
MathDial-SFTTutor Setting=Base in-context2026.040—675.94
SocraticLMTutor Setting=Base in-context2026.040—642.31
TutorRL-7BTutor Setting=Base in-context2026.040—757.84