The "end-to-end" label for LLMs is a misnomer. In practice, they depend on a non-differentiable decoding process that requires laborious, hand-tuning of hyperparameters like temperature and top-p. This paper introduces AutoDeco, a novel architecture that enables truly "end-to-end" generation by learning to control its own decoding strategy. We augment the standard transformer with lightweight heads that, at each step, dynamically predict context-specific temperature and top-p values alongside the next-token logits. This approach transforms decoding into a parametric, token-level process, allowing the model to self-regulate its sampling strategy within a single forward pass.
Through extensive experiments on eight benchmarks, we demonstrate that AutoDeco not only significantly outperforms default decoding strategies but also achieves performance comparable to an oracle-tuned baseline derived from "hacking the test set"-a practical upper bound for any static method. Crucially, we uncover an emergent capability for instruction-based decoding control: the model learns to interpret natural language commands (e.g., "generate with low randomness") and adjusts its predicted temperature and top-p on a token-by-token basis, opening a new paradigm for steerable and interactive LLM decoding.
Why Diffusion Language Models Struggle with Truly Parallel (Non-Autoregressive) Decoding?
The Hong Kong Polytechnic University, Hong Kong, China(香港理工大学);ELLIS Institute Tübingen, Tübingen, Germany(图宾根ELLIS研究所);Max Planck Institute for Intelligent Systems, Tübingen, Germany(智能系统马克斯·普朗克研究所);Tübingen AI Center, Tübingen, Germany(图宾根人工智能中心);University of Surrey, Guildford, United Kingdom(萨里大学);The University of North Carolina at Chapel Hill, Chapel Hill, NC, USA(北卡罗来纳大学教堂山分校)
AdaEAGLE: Optimizing Speculative Decoding via Explicit Modeling of Adaptive Draft Structures
X-LANCE Lab(X-LANCE实验室);Department of Computer Science and Engineering(计算机科学与工程系);MoE Key Lab of Artificial Intelligence(教育部人工智能重点实验室);SJTU AI Institute(上海交通大学人工智能研究院);Shanghai Jiao Tong University(上海交通大学)
CroissantLLM: A Truly Bilingual French-English Language Model
Illuin Technology(伊卢因科技);Unbabel(昂巴贝尔公司);Equall(伊阔尔公司);Diabolocom(迪亚博洛科姆公司);MICS, CentraleSupélec(中央理工学院MICS实验室);Instituto de Telecomunicações, Lisboa(里斯本电信研究所);INESC-ID, Lisboa(里斯本INESC-ID研究所);Instituto Superior Técnico, Universidade de Lisboa(里斯本大学高等技术学院);Sorbonne Université, CNRS, ISIR, Paris(巴黎索邦大学法国国家科学研究中心智能与机器人系统研究所);Imperial College London(伦敦帝国学院);Language Technologies Institute, Carnegie Mellon University(卡内基梅隆大学语言技术研究所)
Self-Introspective Decoding: Alleviating Hallucinations for Large Vision-Language Models
The Hong Kong Polytechnic University(香港理工大学);Hong Kong University of Science and Technology(香港科技大学);Tencent AI Lab(腾讯AI实验室);Huazhong University of Science and Technology(华中科技大学);Tsinghua University(清华大学)
Reference Trustable Decoding: A Training-Free Augmentation Paradigm for Large Language Models
National Engineering Research Center for Multimedia Software(国家多媒体软件工程技术研究中心);School of Computer Science, Wuhan University(武汉大学计算机学院);Department of Computer Science and Engineering, Shanghai Jiao Tong University(上海交通大学计算机科学与工程系)
Decoding the Disaster: Multi-Task Geospatial Reasoning with Vision-Language Models and Crowdsourced Imagery for Disaster Mapping
Shandong University of Science and Technology(山东科技大学);Technical University of Munich(慕尼黑工业大学);Wuhan University(武汉大学);Tsinghua University(清华大学);Helmholtz-Zentrum Dresden-Rossendorf(亥姆霍兹德累斯顿罗森多夫研究中心);University of Iceland(冰岛大学)
Rashid: A Cipher-Based Framework for Exploring In-Context Language Learning
Johns Hopkins University, Center for Language and Speech Processing(约翰霍普金斯大学语言与语音处理中心);MaiNLP, Center for Information and Language Processing, LMU Munich(MaiNLP语言信息处理中心,慕尼黑大学)
Parallelism and Generation Order in Masked Diffusion Language Models: Limits Today, Potential Tomorrow
Zhejiang University(浙江大学);Ant Group(蚂蚁集团);Shanghai Jiao Tong University(上海交通大学);University of Chinese Academy of Social Sciences(中国社会科学院大学);Westlake University(西湖大学)