TY  - RPRT
TI  - Rethinking Fine-Tuning when Scaling Test-Time Compute: Limiting Confidence Improves Mathematical Reasoning
AU  - Feng Chen
AU  - Allan Raventos
AU  - Nan Cheng
AU  - Surya Ganguli
AU  - Shaul Druckmann
PY  - 2025
UR  - https://arxiv.org/abs/2502.07154
ID  - 2502.07154
ER  -