List your company

Compositional Information Seeking on CodeClinic (test)

57.1Level 1 AccuracyGRPO Finetuning
44.41247.7065154.294May 10, 2026

Evaluation results

MethodLinks
GRPO FinetuningMode=GRPO Finetuning2026.0557.136.527.940.522,3762.7
Clinical AutoformalizationMode=Autoformalization2026.0551.241.534.145.323,5692.5
ZeroshotMode=Zeroshot2026.0544.932.826.338.130,4793.1