AI RESEARCH
Self-Distillation for Multi-Token Prediction
arXiv CS.LG
•
ArXi:2603.23911v1 Announce Type: cross As Large Language Models (LLMs) scale up, inference efficiency becomes a critical bottleneck. Multi-Token Prediction (MTP) could accelerate LLM inference by predicting multiple future tokens in parallel. However, existing MTP approaches still face two challenges: limited acceptance rates of MTP heads, and difficulties in jointly