List your company

Question Answering on MMLU Formal Logic (test)

73.67AccuracyMA-PoP
35.845245.665155.48565.3049May 7, 2026

Evaluation results

MethodLinks
MA-PoP2026.0573.67
LLP2026.0572.01
Sparse MADT=32026.0571.43
Sparse MADT=52026.0570.84
Self-Consistency2026.0570.84
Decentr. MADT=22026.0570.63
Decentr. MADT=52026.0569.84
Decentr. MADT=32026.0569.04
Sparse MADT=22026.0566.67
ISP2026.0564.29
Free MADT=52026.0562.7
Majority Voting2026.0561.9
Free MADT=22026.0561.11
Free MADT=32026.0561.11
Centr. MADT=32026.0560.32
Centr. MADT=52026.0560.32
Falcon-7BNumber of Agents (M)=12026.0558.73
Centr. MADT=22026.0556.35
Gemma-9BNumber of Agents (M)=12026.0551.59
Qwen-7BNumber of Agents (M)=12026.0549.05
Llama-8BNumber of Agents (M)=12026.0537.94
Mistral-7BNumber of Agents (M)=12026.0537.3