系统梳理大模型生成的解码方法，提升输出与用户意图对齐效率。
Beyond Tokens: A Survey on Decoding Methods for Large Language and Vision-Language Models
- 提出三种新型解码范式，实现更精准的生成控制。
- 解码方法在不重训练下显著提升生成效率与质量。
- 适合希望优化推理性能的研究者与工程人员。
大型语言模型（LLMs）和大型视觉-语言模型（LVLMs）展现出强大的生成能力，但确保其输出符合用户意图仍具挑战性。现有方法多在训练阶段解决此问题，而推理阶段的解码方法则提供更高效、可扩展的解决方案。解码方法通过引导词元级选择、执行序列级生成或并行生成词元，控制模型输出。本文从近期研究中归纳出三种新兴解码范式，系统综述相关方法，指出当前挑战，并探讨未来研究方向。旨在强调解码方法在效率与效果上的优势，提供实际应用视角。更多论文与资源见 https://github.com/wang2226/Awesome-LLM-Decoding。
原文摘要 · Abstract (English)
Large language models (LLMs) and large vision-language models (LVLMs) have demonstrated impressive generative capabilities, yet ensuring their outputs align with user intent is still challenging. While most existing approaches address this issue at the training stage, inference-time approaches like decoding methods offer a more efficient and scalable solution. Decoding methods control model generation by guiding token-level selection, performing sequence-level generation, or generating tokens in parallel to accelerate the process. In this survey, we identify three emerging paradigms from recent works on decoding methods for LLMs and LVLMs, provide a systematic review of these methods, highlight ongoing challenges, and discuss potential future research directions. Our goal is to underscore the efficiency and effectiveness of decoding methods and offer a practical view of their applications. Paper lists and more resources on decoding methods for LLMs and LVLMs can be found at https://github.com/wang2226/Awesome-LLM-Decoding.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。