Logical Reasoning on RP LP Domain, Npred ≤ 60, 6 < δ ≤ 12
93.5Accuracycot w. L = 64
Evaluation Results
| Method | Links | |
|---|---|---|
| cot w. L = 64Reasoning Strategy=CoT, Number of Layers=642026.05 | 93.5 | |
| cot w. L = 128Reasoning Strategy=CoT, Number of Layers=1282026.05 | 91.8 | |
| cot w. L = 128Reasoning Strategy=CoT, Number of Layers=1282026.05 | 83 | |
| cot w. L = 64Reasoning Strategy=CoT, Number of Layers=642026.05 | 81.6 | |
| direct w. L = 128Reasoning Strategy=Direct, Number of Layers=1282026.05 | 79.8 | |
| cot (baseline)Reasoning Strategy=CoT2026.05 | 79.7 | |
| direct w. L = 128Reasoning Strategy=Direct, Number of Layers=1282026.05 | 78 | |
| cot w. L = 96Reasoning Strategy=CoT, Number of Layers=962026.05 | 77.3 | |
| cot w. L = 48Reasoning Strategy=CoT, Number of Layers=482026.05 | 75.2 | |
| direct (baseline)Reasoning Strategy=Direct2026.05 | 73.5 | |
| direct w. L = 64Reasoning Strategy=Direct, Number of Layers=642026.05 | 71.9 | |
| cot w. L = 32Reasoning Strategy=CoT, Number of Layers=322026.05 | 71.3 | |
| cot w. L = 24Reasoning Strategy=CoT, Number of Layers=242026.05 | 70.1 | |
| direct w. L = 96Reasoning Strategy=Direct, Number of Layers=962026.05 | 66.2 | |
| direct w. L = 48Reasoning Strategy=Direct, Number of Layers=482026.05 | 63.6 | |
| cot w. Nepochs = 60Reasoning Strategy=CoT, N_epochs=602026.05 | 62.5 | |
| cot w. L = 16Reasoning Strategy=CoT, Number of Layers=162026.05 | 61 | |
| cot (baseline)Reasoning Strategy=CoT2026.05 | 60.8 | |
| direct w. L = 24Reasoning Strategy=Direct, Number of Layers=242026.05 | 60.3 | |
| direct w. L = 32Reasoning Strategy=Direct, Number of Layers=322026.05 | 59.7 | |
| cot w. Nepochs = 30Reasoning Strategy=CoT, N_epochs=302026.05 | 58.3 | |
| direct (baseline)Reasoning Strategy=Direct2026.05 | 55.4 | |
| direct w. Nepochs = 60Reasoning Strategy=Direct, N_epochs=602026.05 | 54.8 | |
| direct w. L = 16Reasoning Strategy=Direct, Number of Layers=162026.05 | 53.7 | |
| direct w. Nepochs = 30Reasoning Strategy=Direct, N_epochs=302026.05 | 53.3 |