List your company

Multi-hop QA on 2WikiMultihopQA (pass@1 accuracy)

81.5Pass@1 AccuracySwiR
78.979.57580.2580.925Oct 6, 2025

Evaluation results

MethodLinks
SwiRBackbone=Qwen3-8B2025.1081.5
CoT (Greedy)Backbone=Qwen3-8B, decoding=Greedy2025.1079.5
CoTBackbone=Qwen3-8B2025.1079
Soft ThinkingBackbone=Qwen3-8B2025.1079