List your company

Question Answering on HotpotQA (EM only)

42.4EMPiCA
18.37624.61330.8537.087May 10, 2026

Evaluation results

MethodLinks
PiCABackbone=Qwen2.5-7B Instruct2026.0542.4
TIPSBackbone=Qwen2.5-7B Instruct2026.0542.1
PiCABackbone=Qwen2.5-3B Instruct2026.0540
StepSearchBackbone=Qwen2.5-7B Instruct2026.0538.6
Search-R1Backbone=Qwen2.5-7B Instruct2026.0537
StepSearchBackbone=Qwen2.5-3B Instruct2026.0534.5
ZeroSearchBackbone=Qwen2.5-7B Instruct2026.0532.5
Search-R1Backbone=Qwen2.5-3B Instruct2026.0532.4
TIPSBackbone=Qwen2.5-3B Instruct2026.0531.4
RAGBackbone=Qwen2.5-7B Instruct2026.0528.7
ZeroSearchBackbone=Qwen2.5-3B Instruct2026.0526.7
MT-PPOBackbone=Qwen2.5-7B Instruct2026.0526.5
MT-PPOBackbone=Qwen2.5-3B Instruct2026.0525.5
RAGBackbone=Qwen2.5-3B Instruct2026.0525.1
Search-o1Backbone=Qwen2.5-3B Instruct2026.0524
Search-o1Backbone=Qwen2.5-7B Instruct2026.0519.3