Parallel Token Prediction for Language Models

2025-12-24 · Felix Draxler, Justus Will, Farrin Marouf Sofian, Theofanis Karaletsos, Sameer Singh, Stephan Mandt arxiv

Autoregressive decoding in language models is inherently slow, generating only one token per forward pass. We propose Parallel Token Prediction (PTP), a general-purpose framework for predicting multiple tokens in a single model call. PTP moves the source of randomness from post-hoc sampling to random input variables, making future tokens deterministic functions of those inputs and thus jointly predictable in a single forward pass. We prove that a single PTP call can represent arbitrary dependencies between tokens. PTP is trained by distilling an existing model or through inverse autoregressive training without a teacher. Experimentally, PTP achieves a 2.4x speedup on a diverse-task speculative decoding benchmark. We provide code and checkpoints at https://github.com/mandt-lab/ptp.

📄 PDF Abstract BibTeX arXiv:2512.21323

Code (0)

등록된 구현이 없습니다.

Learning to Parallel: Accelerating Diffusion Large Language Models via Learnable Parallel Decoding

2025-09-29 · Wenrui Bao, Zhiben Chen, Dan Xu, Yuzhang Shang arxiv

Autoregressive decoding in large language models (LLMs) requires $\mathcal{O}(n)$ sequential steps for $n$ tokens, fundamentally limiting inference throughput. Recent diffusion-based LLMs (dLLMs) enable parallel token ge…

Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding

2026-05-14 · Xun Fang, Yunchen Li, Hang Yuan, Zhou Yu arxiv

Discrete diffusion language models improve generation efficiency through parallel token prediction, but standard $X_0$ prediction methods introduce factorization errors by approximating the clean token posterior with ind…

Free Draft-and-Verification: Toward Lossless Parallel Decoding for Diffusion Large Language Models

2025-09-30 · Shutong Wu, Jiawei Zhang arxiv

Diffusion Large Language Models (DLLMs) have emerged as a new paradigm of language modeling beyond autoregressive next-token prediction. Taking advantage of their inherent modeling foundations, DLLMs have the great poten…

Code Generation

Efficient Training-Free Multi-Token Prediction via Embedding-Space Probing

2026-03-18 · Raghavv Goel, Mukul Gagrani, Mingu Lee, Chris Lott arxiv

Large Language Models (LLMs) possess latent multi-token prediction (MTP) abilities despite being trained only for next-token generation. We introduce ESP (Embedding-Space Probing), a simple and training-free MTP method t…

ProPD: Dynamic Token Tree Pruning and Generation for LLM Parallel Decoding

2024-02-21 · Shuzhang Zhong, Zebin Yang, Meng Li, Ruihao Gong 외

Recent advancements in generative large language models (LLMs) have significantly boosted the performance in natural language processing tasks. However, their efficiency is hampered by the inherent limitations in autoreg…