@article{beyondwhatseemsnecessaryhiddengainsfroms, title = {Beyond What Seems Necessary: Hidden Gains from Scaling Training-Time Reasoning Length under Outcome Supervision}, author = {Yihao Xue and Allan Zhang and Jianhao Huang and Amit Sahai and Baharan Mirzasoleiman}, year = {2026}, eprint = {2602.00927}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2602.00927}, }