1 paper
Yunfan Zhou, Ye Zhu, Zhihai Wang +3
Token-level certainty is widely used as a proxy for correctness in LLM training and inference. However, the performance of certainty-based methods depends both on the information i…