List your company

Question Answering on HotpotQA (test) (Accuracy, EM, and F1 Metrics)

37.623EMMHGPO-FoF(os)
-0.984929.0382919.061529.08471Jun 3, 2025Aug 6, 2025Oct 10, 2025Dec 14, 2025Feb 16, 2026Apr 22, 2026Jun 26, 2026

Evaluation results

MethodLinks
MHGPO-FoF(os)Framework=LLM + MAS + MARL2025.0637.62341.51250.858
MHGPO-RRFramework=LLM + MAS + MARL2025.0637.04340.86449.724
R2-Searcher2026.0637—47.6
MHGPO-FoFFramework=LLM + MAS + MARL2025.0636.5740.45949.429
R1-Searcher(PPO)Framework=LLM + RAG (+RL)2025.0635.98237.49347.232
MAPPOFramework=LLM + MAS + MARL2025.0634.4538.21746.4
MHGPO-ISFramework=LLM + MAS + MARL2025.0633.70737.67745.582
AutoRefine2026.0633.6—43.8
R1-Searcher(GRPO)Framework=LLM + RAG (+RL)2025.0632.67336.9245.392
ReSearch2026.0630.8—41.1
Search-R12026.0630.1—39
Qwen2.5-72BFramework=LLM2025.0625.56430.30436.183
ZeroSearch2026.0624.4—33.2
Qwen2.5-72BFramework=LLM + MAS2025.0624.25432.51934.631
Search-o1Framework=LLM + RAG (+RL)2025.0617.9223.02325.702
Search-o12026.0615.5—22.5
RAG2026.0614.2—19.3
Llama3.1-8BFramework=LLM2025.0614.01821.26922.782
Llama3.1-8BFramework=LLM + MAS2025.061420.821.452
Direct Inference2026.0613.7—20.2
COT2026.0613.6—19.7
IRCOT2026.0613.2—19.9
PPOFramework=LLM + RL2025.061.20125.43424.521
GRPOFramework=LLM + RL2025.060.527.64327.421