List your company
ResearchFormerly WizWand

Reasoning on Sudoku (test)

100AccuracyLT
-3.94823.038550.02577.0115Oct 3, 2025Oct 12, 2025Oct 22, 2025Oct 31, 2025Nov 10, 2025Nov 19, 2025Nov 29, 2025
Updated 5mo ago

Evaluation results

MethodLinks
LTSize=6M, Depth T=22025.10100
LTSize=6M, Depth T=32025.10100
CCDDSize=6M, Depth T=22025.10100
CCDDSize=6M, Depth T=32025.10100
MDMSize=6M, Depth T=202025.1099.9
MDMSize=6M, Depth T=22025.1088.6
LlamaSize=13B, Depth T=Seq. Len.2025.1033.8
LlamaSize=7B, Depth T=Seq. Len.2025.1027.1
GPT2 scratchSize=303M, Depth T=Seq. Len.2025.1019.4
GPT2 scratchSize=6M, Depth T=Seq. Len.2025.1016.2
EDITSequence Length=128, 0-shot=true, SFT=true2025.110.161
LLaDASFT=true, Sequence Length=128, 0-shot=true2025.110.114
EDITSequence Length=256, 0-shot=true, SFT=true2025.110.113
LLaDASFT=false, Sequence Length=128, 0-shot=true2025.110.104
LLaDASFT=true, Sequence Length=256, 0-shot=true2025.110.082
EDITSequence Length=512, 0-shot=true, SFT=true2025.110.076
LLaDASFT=false, Sequence Length=256, 0-shot=true2025.110.064
LLaDASFT=false, Sequence Length=512, 0-shot=true2025.110.063
LLaDASFT=true, Sequence Length=512, 0-shot=true2025.110.05