arxiv
PublishedApril 27, 2026 at 4:00 AM
—neutral
Multi-Token Prediction via Self-Distillation
Publisher summary· verbatim
arXiv:2602.06019v2 Announce Type: replace Abstract: Existing techniques for accelerating language model inference, such as speculative decoding, require training auxiliary speculator models and building and deploying complex inference pipelines. We consider a new approach for converting a pretrained
Discussion
No replies yet. Be first.
Related coverage
More from ARXIV
arxivFrom Local to Cluster: A Unified Framework for Causal Discovery with Latent Variables11harxivConsequentialist Objectives and Catastrophe11harxivEgoMAGIC- An Egocentric Video Field Medicine Dataset for Training Perception Algorithms11harxivReCast: Recasting Learning Signals for Reinforcement Learning in Generative Recommendation11hOriginally published on arxiv ↗