List your company
ResearchFormerly WizWand

Logical Reasoning on KK

47.5Test Accuracygemma-3-12b-it
3.40414.85226.337.748May 12, 2026May 15, 2026May 18, 2026May 21, 2026May 24, 2026May 27, 2026May 31, 2026
Updated 5mo ago

Evaluation results

MethodLinks
gemma-3-12b-itBackbone=Gemma 3, RL Type=/, Supervision=/, Environment=/2026.0547.5
wd1Generation Length=5122026.0539.71
StableDRLGeneration Length=5122026.0539.29
GDPOGeneration Length=5122026.0538.86
SPGGeneration Length=5122026.0538.29
wd1Generation Length=2562026.0538
d1Generation Length=5122026.0537.57
StableDRLGeneration Length=2562026.0537.57
AdaBlock-dLLMGeneration Length=5122026.0537.43
Diffu-GRPOGeneration Length=5122026.0537.43
LLaDA-8B-InstructGeneration Length=5122026.0537.29
ESPOGeneration Length=5122026.0537.29
d1Generation Length=2562026.0537
Diffu-GRPOGeneration Length=2562026.0536.86
GDPOGeneration Length=2562026.0536.86
SFTGeneration Length=5122026.0536.57
SPGGeneration Length=2562026.0536.57
ESPOGeneration Length=2562026.0536.14
AdaBlock-dLLMGeneration Length=2562026.0533.71
SimpleSEBackbone=Gemma 3, RL Type=Offline, Supervision=No, Environment=No2026.0533.2
SFTGeneration Length=2562026.0532
gemma-3-4b-itBackbone=Gemma 3, RL Type=/, Supervision=/, Environment=/2026.0531
LLaDA-8B-InstructGeneration Length=2562026.0530
Qwen2.5-14B-InstructBackbone=Qwen 2.5, RL Type=/, Supervision=/, Environment=/2026.0526.4
Qwen2.5-7B-InstructBackbone=Qwen 2.5, RL Type=/, Supervision=/, Environment=/2026.0518.1
SimpleSEBackbone=Qwen 2.5, RL Type=Offline, Supervision=No, Environment=No2026.0517.6
Base + AZR-CoderBackbone=Qwen 2.5, RL Type=Online, Supervision=No, Environment=Yes2026.058.5
Base + AZRBackbone=Qwen 2.5, RL Type=Online, Supervision=No, Environment=Yes2026.055.1