Confidence Before Answering: A Paradigm Shift for Efficient LLM Uncertainty Estimation 文章

ArXiv CS.CL2026-06-04NEWSen作者: Changcheng Li, Jiancan Wu, Hengheng Zhang, Zhengsu Chen, Guo An, Junxiang Qiu, Xiang Wang, Qi Tian

摘要

arXiv:2603.05881v2 Announce Type: replace Abstract: Reliable deployment of large language models (LLMs) requires accurate uncertainty estimation. Existing methods are predominantly answer-first, producing confidence only after generating an answer, which measure the correctness of a specific response and limits practical usability. We study a confidence-first paradigm, where the model outputs its confidence before answering, interpreting this score as the model's probability of answering the question correctly under its current policy. We propose CoCA(Co-optimized Confidence and Answers), a GRPO reinforcement learning framework that jointly optimizes confidence calibration and answer accuracy via segmented credit assignment. By assigning separate rewards and group-relative advantages to confidence and answer segments, CoCA enables stable joint optimization and avoids reward hacking.

相关公司

暂无数据

相关人物

暂无数据

相关产品

暂无数据