List your company

Question Answering on HotpotQA (EM, F1, Avg. iter., AUROC)

45EMAB-RAG
3235.37538.7542.125Jun 27, 2026

Evaluation results

MethodLinks
AB-RAGBackbone=Llama-3.2-3B2026.064558.31.90.644
Static RAGBackbone=Llama-3.2-3B2026.0639.556.3——
Static RAGBackbone=Qwen-1.5B2026.0634.547.3——
AB-RAGBackbone=Qwen-1.5B2026.0634.546.21.090.6
AB-RAGBackbone=Claude Haiku2026.063350.71.360.682
Static RAGBackbone=Claude Haiku2026.0632.551.5——