List your company

Question Answering on TriviaQA single-hop (test)

63.9Average Rate @16GRPO
52.66855.58458.561.416Feb 9, 2026

Evaluation results

MethodLinks
GRPOBackbone=Qwen2.5-7B, Agent Configuration=Multi-Agent & LLM Sharing2026.0263.970.9
Dr. MASBackbone=Qwen2.5-7B, Agent Configuration=Multi-Agent & LLM Non-Sharing2026.0263.472.7
GRPOBackbone=Qwen2.5-7B, Agent Configuration=Single-Agent2026.0263.172.4
Dr. MASBackbone=Qwen2.5-7B, Agent Configuration=Multi-Agent & LLM Sharing2026.0263.171.2
Dr. MASBackbone=Qwen2.5-3B, Agent Configuration=Multi-Agent & LLM Sharing2026.0261.770.1
Dr. MASBackbone=Qwen2.5-3B, Agent Configuration=Multi-Agent & LLM Non-Sharing2026.0261.171.7
GRPOBackbone=Qwen2.5-3B, Agent Configuration=Multi-Agent & LLM Non-Sharing2026.0260.670.8
GRPOBackbone=Qwen2.5-3B, Agent Configuration=Single-Agent2026.0258.168.8
GRPOBackbone=Qwen2.5-3B, Agent Configuration=Multi-Agent & LLM Sharing2026.0257.968.4
GRPOBackbone=Qwen2.5-7B, Agent Configuration=Multi-Agent & LLM Non-Sharing2026.0253.164.4