List your company

Reasoning on e-SNLI

78.44AccuracyExGRPO
51.223258.289165.35572.4209Feb 26, 2026

Evaluation results

MethodLinks
ExGRPOStudent Model=Qwen2.5-7B-Instruct2026.0278.44
RevThinkStudent Model=Qwen2.5-7B-Instruct2026.0273.91
SKDStudent Model=Qwen2.5-7B-Instruct2026.0271.2
AnsAugStudent Model=Qwen2.5-7B-Instruct2026.0270.8
Zero-shotStudent Model=Qwen2.5-7B-Instruct2026.0269.33
ExGRPOStudent Model=Gemma-7B-it2026.0263.21
RevThinkStudent Model=Gemma-7B-it2026.0258.3
SKDStudent Model=Gemma-7B-it2026.0254.4
AnsAugStudent Model=Gemma-7B-it2026.0253.75
Zero-shotStudent Model=Gemma-7B-it2026.0252.27