List your company

Multi-hop QA on FRAMES Wiki18 index (test)

37.5Pass@4Gemma3-12B
11.29218.09624.931.704Jun 15, 2026

Evaluation results

MethodLinks
Gemma3-12BSampling Method=DivInit (DI)2026.0637.5
Gemma3-12BSampling Method=Standard parallel sampling (S)2026.0631
Qwen3-8BSampling Method=DivInit (DI)2026.0630.8
Qwen3-8BSampling Method=Standard parallel sampling (S)2026.0624.8
Qwen3-4BSampling Method=DivInit (DI)2026.0620.4
Qwen3-4BSampling Method=Standard parallel sampling (S)2026.0615.5
Gemma3-4BSampling Method=DivInit (DI)2026.0614.7
Qwen3-1.7BSampling Method=DivInit (DI)2026.0613.6
Qwen3-1.7BSampling Method=Standard parallel sampling (S)2026.0613.1
Gemma3-4BSampling Method=Standard parallel sampling (S)2026.0612.3