Saved in:
Bibliographic Details
Main Authors: Huang, Jianhao, Zhou, Zhanpeng, Xia, Renqiu, Mirzasoleiman, Baharan, Su, Weijie, Huang, Wei
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.11912
Tags: Add Tag
No Tags, Be the first to tag this record!
Table of Contents:
  • While next-token prediction (NTP) has been the standard objective for training language models, it often struggles to capture global structure in reasoning tasks. Multi-token prediction (MTP) has recently emerged as a promising alternative, yet its underlying mechanisms remain poorly understood. In this paper, we study how MTP facilitates reasoning, with a focus on planning. Empirically, we show that MTP consistently outperforms NTP on both synthetic graph path-finding tasks and more realistic reasoning benchmarks, such as Countdown and boolean satisfiability problems. Theoretically, we analyze a simplified two-layer Transformer on a star graph task. We prove that MTP induces a two-stage reverse reasoning process: the model first attends to the end node and then reconstructs the path by tracing intermediate nodes backward. This behavior arises from a gradient decoupling property of MTP, which provides a cleaner training signal compared to NTP. Ultimately, our results highlight how multi-token objectives inherently bias optimization toward robust and interpretable reasoning circuits.