STARS: Segment-level Token Alignment via Rejection Sampling in Large Language Models

M. Atif Quamar, M. Areeb, M. Kuznetsov, M. Ozgur Ozmen, Z. Berkay Celik
September, 2025
Project
Type
1
Publication
Accepted at the Frontiers in Probabilistic Inference Workshop at NeurIPS 2025 and under review at the Transactions of the Association for Computational Linguistics (TACL) journal