List your company

Multi-Hop Question Answering on Musique (F1, Collapse Point)

33.36F1 ScoreCalibAdv
19.382423.011226.6430.2688Apr 20, 2026

Evaluation results

MethodLinks
CalibAdvBackbone=Qwen2.5-7B-Base2026.0433.36—
LLDBackbone=Qwen2.5-7B-Base2026.0431.57—
CalibAdvBackbone=Qwen2.5-3B-Base2026.0430.07—
StepSearchBackbone=Qwen2.5-7B-Base2026.0429.01—
GiGPOBackbone=Qwen2.5-7B-Base2026.0428.5—
Search-R1Backbone=Qwen2.5-3B-Base2026.0428.33292
CalibAdvBackbone=Llama3.2-3B-Instruct2026.0426.77—
SimpleTIRBackbone=Qwen2.5-7B-Base2026.0426.2565
SimpleTIRBackbone=Qwen2.5-3B-Base2026.0425.83310
Search-R1Backbone=Qwen2.5-7B-Base2026.0425.6558
MT-GRPOBackbone=Qwen2.5-7B-Base2026.0423.79—
LLDBackbone=Llama3.2-3B-Instruct2026.0423.7—
SimpleTIRBackbone=Llama3.2-3B-Instruct2026.0421.8554
Search-R1Backbone=Llama3.2-3B-Instruct2026.0419.9240