Inference-time scaling has emerged as an effective way to improve generative models at test time by using a verifier to score and select candidate outputs. A common choice is to employ Multimodal Large Language Models (MLLMs) as verifiers, which can improve performance but introduce substantial inference-time cost. Indeed, diffusion pipelines operate in an autoencoder latent space to reduce computation, yet MLLM verifiers still require decoding candidates to pixel space and re-encoding them into the visual embedding space, leading to redundant and costly operations. In this work, we propose Verifier on Hidden States (VHS), a verifier that operates directly on intermediate hidden representations of Diffusion Transformer (DiT) single-step generators. VHS analyzes generator features without decoding to pixel space, thereby reducing the per-candidate verification cost while improving or matching the performance of MLLM-based competitors. We show that, under tiny inference budgets with only a small number of candidates per prompt, VHS enables more efficient inference-time scaling reducing joint generation-and-verification time by 63.3%, compute FLOPs by 51% and VRAM usage by 14.5% with respect to a standard MLLM verifier, achieving a +2.7% improvement on GenEval at the same inference-time budget.
Logic-Regularized Verifier Elicits Reasoning from LLMs
Department of Computer Science and Technology, East China Normal University(东华大学计算机科学与技术系);Institute of Artificial Intelligence (TeleAI), China Telecom(中国电信人工智能研究院(TeleAI))
Pair-In, Pair-Out: Latent Multi-Token Prediction for Efficient LLMs
Gaoling School of Artificial Intelligence, Renmin University of China(中国人民大学首都人工智能学院);AI Platform, Xiaohongshu Inc.(小红书人工智能平台);University of Electronic Science and Technology of China(电子科技大学);Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所)
CritiqueDriveVLM: From Verifier-Guided Reinforcement Learning to Latent Thought Distillation for Autonomous Driving
State Key Laboratory of Networking and Switching Technology(网络与交换技术国家重点实验室);School of Digital Media & Design Arts(数字媒体与设计艺术学院);Beijing University of Posts and Telecommunications(北京邮电大学)
Reference-Free Rating of LLM Responses via Latent Information
Technical University of Munich, Munich Center for Machine Learning, MDSI(慕尼黑技术大学,慕尼黑机器学习中心,MDSI);National Yang Ming Chiao Tung University(阳明交通大学);Helmholtz Munich(海德堡慕尼黑研究所);Harvard University(哈佛大学)
Dynamical Systems Theory Behind a Hierarchical Reasoning Model
Department of Radiophysics, University of Nizhny Novgorod(尼日尼诺夫戈罗德大学无线电物理系);Huawei Nizhny Novgorod Research Center(华为尼日尼诺夫戈罗德研究中心);Skolkovo Institute of Science and Technology(斯克洛科夫科学与技术研究院)
Competence, Not Accuracy: A Diagnostic for Reference-Free Judge Gates in Skill Optimization
University of California, Los Angeles(加利福尼亚大学洛杉矶分校);Shanghai Jiao Tong University(上海交通大学);Arizona State University(亚利桑那州立大学);Eastern Institute of Technology(东方理工学院)
TapSampling: Inference-Time Sampling with a Task-Progress-Understanding Verifier for Robotic Manipulation
Harbin Institute of Technology, China(哈尔滨工业大学,中国);Harbin Institute of Technology (Weihai) Qingdao Research Institute, China(哈尔滨工业大学(威海)青岛研究院,中国);Iray Technology co., Ltd., Shandong, China(Iray科技有限公司,山东,中国);Tsinghua University, Beijing, China(清华大学,北京,中国)
From Attenuation to Attention: Variational Information Flow Manipulation for Fine-Grained Visual Perception
Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Sciences (ICT/CAS)(智能信息处理重点实验室,计算技术研究所,中国科学院(ICT/CAS));State Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences (ICT/CAS)(人工智能安全国家重点实验室,计算技术研究所,中国科学院(ICT/CAS));University of Chinese Academy of Sciences, Beijing, China(中国科学院大学,北京,中国)
Adaptive Inference-Time Scaling via Early-Step Latent Verification for Image Editing
College of Computer Science and Artificial Intelligence, Fudan University(复旦大学计算机科学与人工智能学院);Microsoft Research Asia(微软亚洲研究院);Institute of Trustworthy Embodied AI, Fudan University(复旦大学可信具身人工智能研究所)