List your company

Navigation Reasoning on Gridworld 10x10 (in-domain)

100Pass@1TSLM
20.9641.486282.52Jan 30, 2026

Evaluation results

MethodLinks
TSLMevaluation_protocol=pass@12026.01100—
Procedure Cloningevaluation_protocol=pass@12026.0199.7—
SCevaluation_protocol=pass@12026.0178.2—
GRPOevaluation_protocol=pass@12026.0124—
Tree-of-Thoughtevaluation_protocol=pass@1002026.01—95