List your company

Multi-Hop Question Answering on Bamboogle (EM, LLM)

49.6EMSearch-R1 (PPO/OF)
818.829.640.4Jun 9, 2026

Evaluation results

MethodLinks
Search-R1 (PPO/OF)Backbone=Qwen3-8B, FT=Full2026.0649.659.2
DACBackbone=Qwen3-8B, FT=LoRA2026.0649.661.6
Search-R1 (GRPO/O)Backbone=Qwen3-8B, FT=Full2026.0648.856.8
Search-R1 (GRPO/OF)Backbone=Qwen3-8B, FT=Full2026.0648.859.2
Search-R1 (PPO/O)Backbone=Qwen3-8B, FT=Full2026.064858.4
DACBackbone=Qwen2.5-7B-Instruct, FT=LoRA2026.0641.652
Search-R1 (PPO/O)Backbone=Qwen2.5-7B-Instruct, FT=Full2026.0640.853.6
Search-R1 (GRPO/O)Backbone=Qwen2.5-7B-Instruct, FT=Full2026.064050.4
Search-R1 (PPO/OF)Backbone=Qwen2.5-7B-Instruct, FT=Full2026.0637.648
IRCoTBackbone=Qwen3-8B, FT=-2026.0636.847.2
Search-R1 (GRPO/OF)Backbone=Qwen2.5-7B-Instruct, FT=Full2026.0635.245.6
Search-o1Backbone=Qwen2.5-7B-Instruct, FT=-2026.0633.641.6
CoTBackbone=Qwen3-8B, FT=-2026.0632.840.8
Search-o1Backbone=Qwen3-8B, FT=-2026.063248
Naive RAGBackbone=Qwen3-8B, FT=-2026.0626.434.4
CoTBackbone=Qwen2.5-7B-Instruct, FT=-2026.0625.630.4
Naive RAGBackbone=Qwen2.5-7B-Instruct, FT=-2026.062428
IRCoTBackbone=Qwen2.5-7B-Instruct, FT=-2026.0615.226.4
DirectBackbone=Qwen3-8B, FT=-2026.0610.412.8
DirectBackbone=Qwen2.5-7B-Instruct, FT=-2026.069.611.2