Not All Tokens Deserve Equal Credit: Counterfactual Sensitivity Credit Reallocation for Long-CoT Reasoning
Open paper
Benchmark
Model
Rank
Results
mathematical-reasoning-on-aime-2025
CSCR (Qwen3-4B)
–
Accuracy: 68.3
mathematical-reasoning-on-aime-2026
CSCR (Qwen3-4B)
–
Accuracy: 67.2
mathematical-reasoning-on-aime24
CSCR (Qwen3-4B)
–
Acc: 74.4
Rank counts only results with a code link.