List your company

Multi-Hop Question Answering on 2WikiMultihopQA (F1/Collapse Point)

56.91F1 ScoreCalibAdv
39.521244.035648.5553.0644Apr 20, 2026

Evaluation results

MethodLinks
CalibAdvBackbone=Qwen2.5-7B-Base2026.0456.91—
LLDBackbone=Qwen2.5-7B-Base2026.0456.17—
Search-R1Backbone=Qwen2.5-7B-Base2026.045358
CalibAdvBackbone=Qwen2.5-3B-Base2026.0452.51—
SimpleTIRBackbone=Qwen2.5-7B-Base2026.0449.9565
Search-R1Backbone=Qwen2.5-3B-Base2026.0449.83292
SimpleTIRBackbone=Qwen2.5-3B-Base2026.0448.75310
LLDBackbone=Llama3.2-3B-Instruct2026.0447.8—
GiGPOBackbone=Qwen2.5-7B-Base2026.0447.79—
CalibAdvBackbone=Llama3.2-3B-Instruct2026.0447.36—
StepSearchBackbone=Qwen2.5-7B-Base2026.0446.94—
MT-GRPOBackbone=Qwen2.5-7B-Base2026.0446.46—
SimpleTIRBackbone=Llama3.2-3B-Instruct2026.0442.8454
Search-R1Backbone=Llama3.2-3B-Instruct2026.0440.1940