Test-Time Prompt Tuning

Ask about this topic

Overview

Test-time prompt tuning studies how to adapt a pretrained model at inference using prompts and unlabeled examples, often when test data differ from training data. Common approaches freeze the model’s main weights and update soft prompts using entropy minimization, consistency across augmented views, or alignment to prototypes and feature distributions. Much of the work targets vision-language models and evaluates accuracy, robustness, and calibration on shifted image benchmarks, while newer applications include graphs, video, LiDAR, and language models. A central tension is that noisy predictions can mislead adaptation or make it overconfident, so methods increasingly filter views, calibrate confidence, or constrain updates while balancing stability, performance, and computational cost.

Latest papers

CardsList

26 papers1 paper in the last four weeks, down 67% on the four weeks before. 0.0% of all new papers.

  1. Test-Time Prompt-Agnostic Decomposition

    Oct 26, 2026Junze Wang, Lei Fan, Dezheng Zhang +4Test-Time AdaptationTest-Time Prompt Tuning

    University of Science and Technology Beijing · University of New South Wales · Sympflux Co., Ltd. +1

  2. FedRAPP: Federated Test-Time Prompt Tuning for Vision-Language Models with Reliability-Aware Prompt Personalization

    Oct 16, 2026Syed Md. Ahnaf Hasan, Md Akil Raihan Iftee, Sajib Mistry +2Prompt Learning for VLMsTest-Time Prompt Tuning

    Center for Computational & Data Sciences (CCDS), Independent University, Bangladesh · Rochester Institute of Technology, USA · Curtin University, Australia

  3. Balanced Prompt Adaptation against Entropy-Induced Collapse for Test-Time Binary Segmentation

    Sep 18, 2026Zhengshan Wang, Joshua Charles Webster-Ford, Yifei Tian +3Test-Time AdaptationImage Segmentation

    Shenzhen University, Shenzhen, China

  4. Bridging the Confidence Gap: Temperature Scaling for Calibrating Test-Time Prompt Tuning

    Sep 15, 2026Yuwei Liang, Jian Liang, Dapeng Hu +2Post-Hoc CalibrationTest-Time Adaptation of VLMs

    School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China · NLPR & MAIS, Institute of Automation, Chinese Academy of Sciences, Beijing, China · Independent Researcher

  5. Rethinking the Test-Time Prompt Tuning Objective from the Perspective of Calibration

    Aug 31, 2026Jungwon Choi, Hyeonseo Jang, Kibok Lee +1VLM CalibrationUncertainty Calibration

    Department of Computer Science and Engineering, Chung-Ang University · Department of Statistics and Data Science, Yonsei University

  6. Fairness-Aware Test-Time Prompt Tuning

    Aug 26, 2026Yoann Launay, Parameswaran Kamalaruban, Tom Kempton +2Test-Time Adaptation of VLMsAlgorithmic Fairness

    University of Cambridge · Risk and Security AI Lab, Visa Inc. · University of Manchester

  7. PromptResponse: Optimizing Prompts for LLM Coding Tasks

    Aug 21, 2026Erik Thureck, Robert Kühnen, Tim JacobowitzLLM PromptingLLM Code Generation

    HUMBOLDT-UNIVERSITÄT ZU BERLIN MATHEMATISCH-NATURWISSENSCHAFTLICHE FAKULTÄT INSTITUT FÜR INFORMATIK HUMAN-COMPUTER INTERACTION LAB

  8. Robustifying Vision-Language Models via Test-Time Prompt Adaptation

    Jul 10, 2026Xingyu Zhu, Huanshen Wu, Shuo Wang +4VLM RobustnessVLM Adaptation

    National University of Singapore · University of Science and Technology of China · The Hong Kong University of Science and Technology

  9. C-GAP: Class-Aware and Online Prompting Improves Vision-Language Models on Imbalanced Classes

    Jul 10, 2026Francis Fernandez, Arash Jahangiri, Salimeh SekehOpen-Vocabulary Object DetectionObject Detection

    San Diego State University

  10. USE: A Unified Self-Ensembling Framework for Test-Time Prompt Tuning

    Jul 4, 2026Siru Jiang, Jian Liang, Ran He +1Ensemble LearningVLM Adaptation

    School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences · NLPR & MAIS, Institute of Automation, Chinese Academy of Sciences · School of Artificial Intelligence, University of Chinese Academy of Sciences +1

  11. No Adaptation Without Observation: Observability-Constrained Test-Time Prompt Tuning for LiDAR Semantic Segmentation

    Jun 29, 2026Linlian Jiang, Wentao Ju, Sadman Rakib Pinon +4Test-Time Adaptation3D Semantic Segmentation

    Concordia University · Mila – Quebec AI Institute · University of Toronto

  12. PromptDyG: Test-Time Prompt Adaptation on Dynamic Graphs

    Jun 22, 2026Guoguo Ai, Chaoxi Niu, Hui Yan +3Continual Test-Time AdaptationDynamic Graph Learning

    School of Computer Science and Engineering, Nanjing University of Science and Technology, China · School of Computing and Information Systems, Singapore Management University, Singapore · Faculty of Data Science, City University of Macau, Macau, China +3

  13. Beyond Prompt and Parser Engineering: Soft-Prompt Tuning for Automated LLM Evaluation

    Jun 10, 2026Selen Erkan, Bastian Boll, Kristian Kersting +2LLM EvaluationInstruction Following

    Aleph Alpha Research Lab1141 · TU Darmstadt Hessian.AI Lab1141

  14. EEVEE: Towards Test-time Prompt Learning in the Real World for Self-Improving Agents

    Jun 9, 2026Weixian Xu, Shilong Liu, Mengdi WangContinual Learning for LLM AgentsLLM Agent Self-Improvement

    Shanghai Jiao Tong University · Princeton University

  15. SS-TPT: Stability and Suitability-Guided Test-Time Prompt Tuning for Adversarially Robust Vision-Language Models

    Jun 5, 2026Sunoh Kim, Daeho UmVLM RobustnessVLM Adaptation

    Dankook University, Yongin, South Korea · University of Seoul, Seoul, South Korea

  16. From Demonstrations to Rewards: Test-Time Prompt Optimization for VLM Reward Models

    May 22, 2026Christian Gumbsch, Leonardo Barcellona, Lennard Schünemann +7RL for RoboticsMultimodal Reward Modeling

    University of Amsterdam · 2Catholic University of Leuven · 3Toyota Research Institute +1

  17. Self-Improving In-Context Learning

    May 22, 2026Baturay Saglam, Dionysis KalogeriasLanguage Model CalibrationIn-Context Learning

    Department of Electrical and Computer Engineering Yale University

  18. Towards Fine-Grained Robustness: Attention-Guided Test-Time Prompt Tuning for Vision-Language Models

    May 19, 2026Jia-Wei Hai, Yijun Wang, Xiu-Shen WeiVLM RobustnessAdversarial Attacks on VLMs

    School of Computer Science and Engineering, and Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications, Southeast University, China. · School of Intelligence Science and Engineering, Southeast University, China.

  19. TAME: Test-Time Adversarial Prompt Tuning via Mixture-of-Experts for Vision-Language Models

    May 17, 2026Xin Wang, Yixu Wang, Jiaming Zhang +6Adversarial Attacks on VLMsTest-Time Adaptation

  20. GFMate: Empowering Graph Foundation Models with Test-time Prompt Tuning

    May 14, 2026Yan Jiang, Ruihong Qiu, Zi HuangDomain AdaptationGraph Domain Adaptation

    School of Electrical Engineering and Computer Science, The University of Queensland, Brisbane, Queensland, Australia

  21. Improving Calibration in Test-Time Prompt Tuning for Vision-Language Models via Data-Free Flatness-Aware Prompt Pretraining

    Apr 30, 2026Hyeonseo Jang, Jaebyeong Jeon, Joong-Won Hwang +1Deep Learning OptimizationVision-Language Models

    1Yonsei University · 2ETRI

  22. Dual-Modality Anchor-Guided Filtering for Test-time Prompt Tuning

    Apr 14, 2026Jungwon Choi, Eunwoo KimVLM RobustnessVLM Adaptation

    School of Computer Science and Engineering, Chung-Ang University

  23. P-Flow: Prompting Visual Effects Generation

    Mar 23, 2026Rui Zhao, Mike Zheng ShouControllable Video GenerationT2V Generation

    Show Lab, National University of Singapore

  24. Geometry-Aware Semantic Reasoning for Training Free Video Anomaly Detection

    Mar 10, 2026Ali Zia, Usman Ali, Muhammad Umer Ramzan +3Training-Free Video Anomaly DetectionVideo Anomaly Detection

    School of Computing, Engineering & Mathematical Sciences, La Trobe University, Melbourne, Australia · School of Engineering and Applied Sciences, GIFT University, Gujranwala 52250, Pakistan

  25. FOZO: Forward-Only Zeroth-Order Prompt Optimization for Test-Time Adaptation

    Mar 5, 2026Xingyu Wang, Tao WangTest-Time AdaptationContinual Test-Time Adaptation

    College of Computer Science, Sichuan University

  26. Context Tuning for In-Context Optimization

    Jul 6, 2025Jack Lu, Ryan Teehan, Zhenbang Yang +1Few-Shot LearningIn-Context Learning

    Agentic Learning AI Lab, New York University