List your company

Logical Reasoning on Zebra

36.5Pass@1 AccuracySEC
-0.3169.24218.828.358Jun 23, 2026

Evaluation results

MethodLinks
SECBackbone=Llama3.2-3B2026.0636.5
TACBackbone=Llama3.2-3B2026.0635.8
TACBackbone=Qwen3-1.7B2026.0634.7
M2OBackbone=Llama3.2-3B2026.0630.9
RandomBackbone=Llama3.2-3B2026.0630.5
M2OBackbone=Qwen3-1.7B2026.0628.5
SECBackbone=Qwen3-1.7B2026.0626.7
RandomBackbone=Qwen3-1.7B2026.0624.9
BaseBackbone=Llama3.2-3B2026.061.4
BaseBackbone=Qwen3-1.7B2026.061.1