Inference-Time Language Model Alignment

Follow this topic Ask about this topic

The field studies how to guide a language model toward desired behavior at generation time, often without changing its weights. Approaches include prompt and persona conditioning, activation or logit steering, reward-guided decoding, and sampling or reranking methods such as Best-of-N. Researchers apply these methods to safety, helpfulness, reasoning, personalization, and conflicting preferences, while developing theoretical guarantees and more efficient ways to evaluate them. A central challenge is balancing gains in alignment against compute cost, reward-model errors, loss of useful capabilities, and the difficulty of transferring results across tasks and conversations.

Latest papers

78 papers13 papers in the last four weeks, up 8% on the four weeks before. 0.1% of all new papers.

Towards Clinically Faithful Medical Image Captioning via Enhanced Vision-Language Alignment
Aug 20, 2026Yunseo Lee, Hyun Jun Kim, Heeseung Shin +1Medical VLMs Vision-Language Alignment
Department of Statistics and Data Science, Chung-Ang University, 84 Heukseok-ro, Dongjak-gu, Seoul 06974, Republic of Korea · Department of Smart Cities, Chung-Ang University, 84 Heukseok-ro, Dongjak-gu, Seoul 06974, Republic of Korea · Department of Applied Statistics, Chung-Ang University, 84 Heukseok-ro, Dongjak-gu, Seoul 06974, Republic of Korea
Truncate Bad, Upweight Good: BoN-Style Distillation via Rank-Based Classification
Aug 20, 2026Yarin Bar, Yaniv RomanoLanguage Model Distillation Knowledge Distillation
Department of Electrical and Computer Engineering Technion–Israel Institute of Technology · Department of Computer Science Technion–Israel Institute of Technology
Towards a theory of inference-time alignment with unknown rewards
Aug 15, 2026Steve Hanneke, Hongao Wang, Mingyue XuInference-Time Language Model Alignment Statistical Learning Theory
Purdue University
Weightless Fine-Tuning: Personalizing LLMs via Logit-Space Transport
Aug 11, 2026Bohan Zhang, Anqi Ni, Yixin Wang +1Supervised Fine-Tuning LLM Personalization
University of Michigan · University of Chicago
CIDER: A Dataset of Contextual Disclosure Boundaries for Privacy Preference Alignment
Aug 10, 2026Bingcan Guo, Eryue Xu, Jijie Zhou +2Preference Alignment Inference-Time Language Model Alignment
University of Washington · UIUC · Northeastern University
Deployable Per-Instance Multi-Layer Activation Steering for Large Language Models
Aug 9, 2026Muhammad Faishal Adly Nelwan, Alfan Farizki WicaksonoLanguage Model Steering Controllable Text Generation
Faculty of Computer Science Universitas Indonesia
Beyond a Global Norm: Personalizing Toxicity Sensitivity in Language Models Without Retraining
Jul 25, 2026Rares A. C. Diaconescu, Iulia Slanina, Alina Florea +5LLM Alignment LLM Safety Alignment
Delft University of Technology
Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs
Jul 21, 2026Alexander ManevMultilingual Language Model Evaluation LLM Alignment
Technical University of Munich
Best-of-Better- $N$ : Generating Pre-Aligned Responses with In-Context Learning
Jul 3, 2026Eric Lei, Hsiang Hsu, Chun-Fu ChenLLM Alignment Best-of-N Sampling
JPMorganChase Global Technology Applied Research
Safe Inference-Time Alignment via Lagrangian Reward Augmentation
Jul 2, 2026Yaswanth Chittepu, Ativ Joshi, Sohini Chintala +1LLM Alignment LLM Safety Alignment
University of Massachusetts Amherst · Independent Researcher
Towards Spec Learning: Inference-Time Alignment from Preference Pairs
Jun 22, 2026Dhriti Krishnan, Tejas Goyal, Jaromir SavelkaLLM Alignment Prompt Optimization
Department of Computer Science Carnegie Mellon University
PG-MAP: Joint MAP Optimization for Inference-Time Alignment of Diffusion and Flow-Matching Models
Jun 22, 2026Ruolan Sun, Pawel PolakDiffusion Model Alignment Inference-Time Language Model Alignment
Stony Brook University
ALIGNBEAM : Inference-Time Alignment Transfer via Cross-Vocabulary Logit Mixing
Jun 10, 2026Chirag Chawla, Pratinav Seth, Vinay Kumar SankarapuLLM Alignment LLM Safety Alignment
Lexsi Labs
Gradient-Guided Reward Optimization for Inference-time Alignment
Jun 8, 2026Hankun Lin, Ruqi ZhangReward Hacking LLM Alignment
Department of Computer Science, Purdue University
Inference-Time Vulnerability Beyond Shallow Safety: Alignment Along Generation Trajectories
Jun 3, 2026Kyungmin Park, Taesup KimAdversarial Attacks on LLMs LLM Safety Alignment
Hankuk University of Foreign Studies · Seoul National University
Effort as Ceiling, Not Dial: Reasoning Budget Does Not Modulate Cognitive Cost Alignment Between Humans and Large Reasoning Models
May 16, 2026Yueqing Hu, Tianhong WangComputational Cognitive Modeling Inference-Time Scaling
Institute of Neuroscience, Chinese Academy of Sciences, Shanghai, China · School of Philosophy, Anhui University, Hefei, China
Temper and Tilt Lead to SLOP: Reward Hacking Mitigation with Inference-Time Alignment
May 13, 2026Ye Wang, Jing Liu, Toshiaki Koike-AkinoReward Hacking LLM Alignment
Mitsubishi Electric Research Laboratories (MERL) · Broadway, Cambridge, MA 02139, USA
Training-Free Cultural Alignment of Large Language Models via Persona Disagreement
May 11, 2026Huynh Trung Kiet, Dao Sy Duy Minh, Tuan Nguyen +5LLM Alignment Cultural Alignment
Faculty of Information and Technology, University of Science, Vietnam National University, Ho Chi Minh City, Vietnam. · Department of Computer Science, University of Warwick, Coventry, United Kingdom. · School of Computing, Engineering and Digital Technologies, Teesside University, United Kingdom.
One Refiner to Unlock Them All: Inference-Time Reasoning Elicitation via Reinforcement Query Refinement
Apr 28, 2026Yixiao Zhou, Dongzhou Cheng, zhiliang wu +3LLM Alignment RL for Language Model Reasoning
Zhejiang University · Shanghai Innovation Institute · 1Zhejiang University, 2Shanghai Innovation Institute +9
Pref-CTRL: Preference Driven LLM Alignment using Representation Editing
Apr 26, 2026Imranul Ashrafi, Inigo Jauregi Unanue, Massimo PiccardiLLM Alignment Representation Engineering
University of Technology Sydney, Australia
To Intervene or Not: Guiding Inference-time Alignment with Probabilistic Model Blending
Apr 22, 2026Jin Gan, Xin Li, Jun LuoLLM Alignment Inference-Time Language Model Alignment
College of Computing and Data Science, Nanyang Technological University, Singapore
On the Rejection Criterion for Proxy-based Test-time Alignment
Apr 17, 2026Ayoub Hammal, Pierre Zweigenbaum, Caio CorroLLM Alignment Inference-Time Language Model Alignment
Université Paris-Saclay, CNRS, LISN · INSA Rennes, IRISA, CNRS, Université de Rennes
Hierarchical Alignment: Enforcing Hierarchical Instruction-Following in LLMs through Logical Consistency
Apr 10, 2026Shu Yang, Zihao Zhou, Di Wang +1LLM Alignment Inference-Time Language Model Alignment
King Abdullah University of Science and Technology (KAUST) · University of Liverpool · University of Edinburgh
Where to Steer: Input-Dependent Layer Selection for Steering Improves LLM Alignment
Apr 4, 2026Soham Gadgil, Chris Lin, Su-In LeeLanguage Model Steering Inference-Time Language Model Alignment
University of Washington
Structural Rigidity and the 57-Token Predictive Window: A Physical Framework for Inference-Layer Governability in Large Language Models
Apr 4, 2026Gregory M. RuddellLanguage Model Safety Evaluation Inference-Time Language Model Alignment
SnailSafe.ai — Carson City, Nevada
Deliberative Alignment is Deep, but Uncertainty Remains: Inference time safety improvement in reasoning via attribution of unsafe behavior to base model
Apr 1, 2026Pankayaraj Pathmanathan, Furong HuangLLM Alignment Inference-Time Language Model Alignment
University of Maryland College Park
DSPA: Dynamic SAE Steering for Data-Efficient Preference Alignment
Mar 23, 2026James Wedgwood, Aashiq Muhamed, Mona T. Diab +1Preference Alignment Inference-Time Language Model Alignment
Carnegie Mellon University
TARo: Token-level Adaptive Routing for LLM Test-time Alignment
Mar 19, 2026Arushi Rai, Qiang Zhang, Hanqing Zeng +5Inference-Time Language Model Alignment LLM Reasoning
Meta · University of Pittsburgh · University of California, Berkeley
MOSAIC: Composable Safety Alignment with Modular Control Tokens
Mar 17, 2026Jingyu Peng, Hongyu Chen, Jiancheng Dong +5LLM Safety Alignment Inference-Time Language Model Alignment
University of Science and Technology of China · City University of Hong Kong · Baidu Inc +1
Test-time RL alignment exposes task familiarity artifacts in LLM benchmarks
Mar 13, 2026Kun Wang, Reinhard HeckelLLM Evaluation Inference-Time Language Model Alignment
School of Computation, Information and Technology, Technical University of Munich · Munich Center for Machine Learning

Aug 20, 2026J/K move · Enter open · S save

Towards Clinically Faithful Medical Image Captioning via Enhanced Vision-Language Alignment

Yunseo Lee, Hyun Jun Kim, Heeseung Shin, Changwon Lim

Department of Statistics and Data Science, Chung-Ang University, 84 Heukseok-ro, Dongjak-gu, Seoul 06974, Republic of Korea · Department of Smart Cities, Chung-Ang University, 84 Heukseok-ro, Dongjak-gu, Seoul 06974, Republic of Korea · Department of Applied Statistics, Chung-Ang University, 84 Heukseok-ro, Dongjak-gu, Seoul 06974, Republic of Korea

Medical VLMs Vision-Language Alignment Faithfulness in VLMs Inference-Time Language Model Alignment

Open paper AskPDF arXiv

Medical image captioning is a technique that accelerates early-stage diagnostic workflows and enhances the interpretability of medical diagnostic AI systems. However, unlike general image captioning, clinically reliable captioning remains challenging due to grayscale-based modalities, subtle anatomical cues, specialized medical phrasing, and variations in data quality. Despite recent advances in large vision-language models, fluent outputs do not necessarily guarantee sufficient alignment with clinical concept spaces or evaluation criteria. To address this issue, we propose a framework that strengthens clinical alignment by separating and enhancing training-time alignment and inference-time alignment. We build a medical image captioning pipeline that integrates single/dual vision encoders based on BioMedCLIP and SigLIP2, a Q-Former, and a LLaMA-based decoder, and examine the contribution of auxiliary learning for UMLS concept/type prediction. At inference, we apply single-embedding-based reranking to select the best caption among candidates, while at training we introduce MedPAIR-SCST, which combines clinically relevant rewards to shift the generative distribution toward improved clinical alignment. Our experiments show that complementary visual representations with a multi-encoder design and concept-level auxiliary learning help preserve clinically meaningful information. Furthermore, inference-time reranking provides a practical way to improve semantic and clinical alignment without additional training, whereas MedPAIR-SCST goes beyond selection by directly improving the model's distribution to generate more consistent and clinically grounded captions. These findings suggest that jointly leveraging selection-based alignment and reinforcement-learning-based alignment can promote more trustworthy medical image captioning even in data-constrained settings.