List your company

Tutor Robustness Evaluation on Finetuned Adv. Agent

1Student Information LeakageSocraticLM
0.961.231.51.77Apr 20, 2026

Evaluation results

MethodLinks
SocraticLMTutor Setting=Base in-context2026.04113.52702.61
MathDial-SFTTutor Setting=Base in-context2026.0428.21706.1
TutorRL-7BTutor Setting=Base in-context2026.04210.94838.65