Propositional Logic Reasoning on ProsQA (val)
93.37AccuracyCoT2
Evaluation Results
| Method | Links | |
|---|---|---|
| CoT2Backbone=4-layer, 4-head GPT2 (d=32), Budget=Full2025.05 | 93.37 | |
| COCONUTBackbone=4-layer, 4-head GPT2 (d=32)2025.05 | 90.03 | |
| Discrete CoTBackbone=4-layer, 4-head GPT2 (d=32), Budget=B=12025.05 | 68.5 | |
| Discrete no-CoTBackbone=4-layer, 4-head GPT2 (d=32)2025.05 | 54.91 |