List your company

Multi-Hop Question Answering on 2WikiMultihopQA (EM, LLM)

46EMDAC
16.77624.36331.9539.537Jun 9, 2026

Evaluation results

MethodLinks
DACBackbone=Qwen3-8B, FT=LoRA2026.064652.8
Search-R1 (GRPO/OF)Backbone=Qwen3-8B, FT=Full2026.0643.950.1
Search-R1 (PPO/OF)Backbone=Qwen3-8B, FT=Full2026.0642.247.9
Search-R1 (GRPO/O)Backbone=Qwen3-8B, FT=Full2026.0640.546.3
Search-R1 (PPO/O)Backbone=Qwen3-8B, FT=Full2026.0638.243.6
DACBackbone=Qwen2.5-7B-Instruct, FT=LoRA2026.0635.441.4
Search-R1 (GRPO/O)Backbone=Qwen2.5-7B-Instruct, FT=Full2026.0634.839.5
Search-R1 (GRPO/OF)Backbone=Qwen2.5-7B-Instruct, FT=Full2026.0634.740.5
Search-R1 (PPO/O)Backbone=Qwen2.5-7B-Instruct, FT=Full2026.0634.440
Search-R1 (PPO/OF)Backbone=Qwen2.5-7B-Instruct, FT=Full2026.0632.938.8
Search-o1Backbone=Qwen3-8B, FT=-2026.0631.740.9
IRCoTBackbone=Qwen3-8B, FT=-2026.0626.334.4
Search-o1Backbone=Qwen2.5-7B-Instruct, FT=-2026.062632.9
CoTBackbone=Qwen3-8B, FT=-2026.0625.527.6
DirectBackbone=Qwen3-8B, FT=-2026.0625.227.1
DirectBackbone=Qwen2.5-7B-Instruct, FT=-2026.0624.625.8
CoTBackbone=Qwen2.5-7B-Instruct, FT=-2026.0621.222.8
IRCoTBackbone=Qwen2.5-7B-Instruct, FT=-2026.0619.531.1
Naive RAGBackbone=Qwen3-8B, FT=-2026.0619.322.7
Naive RAGBackbone=Qwen2.5-7B-Instruct, FT=-2026.0617.921.2