List your company

Multi-Hop Question Answering on MuSiQue (Exact Match %)

22.6Exact Match (%)StepSearch
5.3369.81814.318.782Jul 7, 2026

Evaluation results

MethodLinks
StepSearchBackbone=Qwen2.5-7B-Instruct, Type=Chain-based2026.0722.6
IGRPOBackbone=Qwen2.5-7B-Instruct, Type=Tree-based2026.0720.2
GiGPOBackbone=Qwen2.5-7B-Instruct, Type=Chain-based2026.0718.9
ZeroSearchBackbone=Qwen2.5-7B-Instruct, Type=Chain-based2026.0718.4
Tree-GRPOBackbone=Qwen2.5-7B-Instruct, Type=Tree-based2026.0717.5
StepSearchBackbone=Qwen2.5-3B-Instruct, Type=Chain-based2026.0717.4
IGPOBackbone=Qwen2.5-7B-Instruct, Type=Chain-based2026.0716.8
Search-R1Backbone=Qwen2.5-7B-Instruct, Type=Chain-based2026.0714.6
IGRPOBackbone=Qwen2.5-3B-Instruct, Type=Tree-based2026.0714.1
GiGPOBackbone=Qwen2.5-3B-Instruct, Type=Chain-based2026.0712.6
IGPOBackbone=Qwen2.5-3B-Instruct, Type=Chain-based2026.0712.3
Tree-GRPOBackbone=Qwen2.5-3B-Instruct, Type=Tree-based2026.0711.9
Search-R1Backbone=Qwen2.5-3B-Instruct, Type=Chain-based2026.0710.3
ZeroSearchBackbone=Qwen2.5-3B-Instruct, Type=Chain-based2026.079.8
R1-InstructBackbone=Qwen2.5-3B-Instruct, Type=RL Training2026.077.2
R1-InstructBackbone=Qwen2.5-7B-Instruct, Type=RL Training2026.076