List your company

Multi-hop QA on HotpotQA (HpQA) Wiki18 index (test)

59.1Pass@4Gemma3-12B
39.23644.39349.5554.707Jun 15, 2026

Evaluation results

MethodLinks
Gemma3-12BSampling Method=DivInit (DI)2026.0659.1
Qwen3-8BSampling Method=DivInit (DI)2026.0657
Gemma3-12BSampling Method=Standard parallel sampling (S)2026.0654.9
Qwen3-4BSampling Method=DivInit (DI)2026.0653.2
Qwen3-8BSampling Method=Standard parallel sampling (S)2026.0650.4
Gemma3-4BSampling Method=DivInit (DI)2026.0649.2
Qwen3-1.7BSampling Method=DivInit (DI)2026.0643.8
Qwen3-1.7BSampling Method=Standard parallel sampling (S)2026.0642.9
Qwen3-4BSampling Method=Standard parallel sampling (S)2026.0641.9
Gemma3-4BSampling Method=Standard parallel sampling (S)2026.0640