List your company

Tool-augmented Question Answering on HotpotQA (Accuracy, Delta)

61AccuracyEPC-AW
47.823251.244154.66558.0859May 22, 2026

Evaluation results

MethodLinks
EPC-AWBackbone=Qwen3-Coder-30B2026.0561—
RollbackBackbone=Qwen3-Coder-30B2026.0556.334.67
RetryBackbone=Qwen3-Coder-30B2026.0551.339.67
No-Repair (AgentFlow)Backbone=Qwen3-Coder-30B2026.0548.3312.67