List your company

Multi-hop Question Answering on HotpotQA (test) (avg@16/pass@16)

44Avg@16Dr. MAS
23.61628.90834.239.492Feb 9, 2026

Evaluation results

MethodLinks
Dr. MASBackbone=Qwen2.5-7B, Agent Configuration=Multi-Agent & LLM Non-Sharing2026.024457.5
GRPOBackbone=Qwen2.5-7B, Agent Configuration=Single-Agent2026.024355.1
Dr. MASBackbone=Qwen2.5-7B, Agent Configuration=Multi-Agent & LLM Sharing2026.0242.556
GRPOBackbone=Qwen2.5-7B, Agent Configuration=Multi-Agent & LLM Sharing2026.0240.355
Dr. MASBackbone=Qwen2.5-3B, Agent Configuration=Multi-Agent & LLM Non-Sharing2026.0235.351.1
Dr. MASBackbone=Qwen2.5-3B, Agent Configuration=Multi-Agent & LLM Sharing2026.0233.351.2
GRPOBackbone=Qwen2.5-3B, Agent Configuration=Multi-Agent & LLM Sharing2026.0232.548
GRPOBackbone=Qwen2.5-3B, Agent Configuration=Multi-Agent & LLM Non-Sharing2026.0232.545.2
GRPOBackbone=Qwen2.5-3B, Agent Configuration=Single-Agent2026.0231.840.9
GRPOBackbone=Qwen2.5-7B, Agent Configuration=Multi-Agent & LLM Non-Sharing2026.0224.436.2