List your company

Multi-Hop Question Answering on HotpotQA in-domain (val test)

49.9Exact Match (EM)Search-R2
7.57218.56129.5540.539Feb 3, 2026

Evaluation results

MethodLinks
Search-R2Backbone=Qwen2.5-32B2026.0249.9
Search-R1Backbone=Qwen2.5-32B2026.0243.3
Search-R2Backbone=Qwen3-8B2026.0241.2
Search-R2Backbone=Qwen2.5-7B2026.0239
Search-R1Backbone=Qwen3-8B2026.0237.2
Rejection SamplingBackbone=Qwen2.5-7B2026.0233.1
Search-R1Backbone=Qwen2.5-7B2026.0232.6
RAGBackbone=Qwen2.5-7B2026.0229.9
R1-baseBackbone=Qwen2.5-7B2026.0224.2
R1-instructBackbone=Qwen2.5-7B2026.0223.7
SFTBackbone=Qwen2.5-7B2026.0221.7
Search-o1Backbone=Qwen2.5-7B2026.0218.7
Direct InferenceBackbone=Qwen2.5-7B2026.0218.3
IRCoTBackbone=Qwen2.5-7B2026.0213.3
CoTBackbone=Qwen2.5-7B2026.029.2