Reasoning failure prediction and recovery on CRUXEval (L3)
74Accuracythought-tree-based classifier
Evaluation results
| Method | Links | ||
|---|---|---|---|
| thought-tree-based classifierTarget Reasoning Model=QwQ, Evaluation Protocol=Generalization from L1 training2026.04 | 74 | 27 | |
| thought-tree-based classifierTarget Reasoning Model=R1, Evaluation Protocol=Generalization from L1 training2026.04 | 73 | 15 |