TestDG: Test-time Domain Generalization for Continual Test-time Adaptation
Not on the map. Why?
Abstract (arXiv)
This paper studies continual test-time adaptation (CTTA), the task of adapting a model to constantly changing unseen domains in testing while preserving previously learned knowledge. Existing CTTA methods mostly focus on adaptation to the current test domain only, overlooking generalization to arbitrary test domains a model may face in the future. To tackle this limitation, we present a novel online test-time domain generalization framework for CTTA, dubbed TestDG. TestDG aims to learn features invariant to both current and previous test domains on the fly during testing, improving the potential for effective generalization to future domains. To this end, we propose a new model architecture and a test-time adaptation strategy dedicated to learning domain-invariant features, along with prototype selection and update mechanisms for effectively managing information from previous test domains. TestDG achieved state of the art on four public CTTA benchmarks. Moreover, it showed superior generalization to unseen test domains.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 79 of its 92 references that are in the map. First posted on arXiv 7 Apr 2025.
Builds on
79 of 92 references.
References · 79
- 2022Training language models to follow instructions with human feedbackcited by 15,971
- 2017Proximal Policy Optimization Algorithmscited by 15,215
- 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
- 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
- 2017Deep reinforcement learning from human preferencescited by 3,912
- 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
- 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
- 2020Learning to summarize from human feedbackcited by 2,588
- 2023A Survey of Large Language Modelscited by 2,576
- 2022Emergent Abilities of Large Language Modelscited by 2,512
- 2023Jailbroken: How Does LLM Safety Training Fail?cited by 1,651
- 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,420
38 more
- 2022Ignore Previous Prompt: Attack Techniques For Language Modelscited by 832
- 2023AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Modelscited by 785
- 2023Tree of Attacks: Jailbreaking Black-Box LLMs Automaticallycited by 672
- 2024How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMscited by 573
- 2023Jailbreaking ChatGPT via Prompt Engineering: An Empirical Studycited by 498
- 2023"Do Anything Now": Characterizing and Evaluating In-The-Wild Jailbreak Prompts on Large Language Modelscited by 489
- 2024Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attackscited by 486
- 2023Catastrophic Jailbreak of Open-source LLMs via Exploiting Generationcited by 451
- 2023Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrationscited by 431
- 2023GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Ciphercited by 423
- 2023FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Promptscited by 398
- 2023A Survey on Large Language Model (LLM) Security and Privacy: The Good, the Bad, and the Uglycited by 387
- 2023DeepInception: Hypnotize Large Language Model to Be Jailbreakercited by 367
- 2023Multi-step Jailbreaking Privacy Attacks on ChatGPTcited by 333
- 2023Visual Adversarial Examples Jailbreak Aligned Large Language Modelscited by 328
- 2023Low-Resource Languages Jailbreak GPT-4cited by 309
- 2023On Evaluating Adversarial Robustness of Large Vision-Language Modelscited by 284
- 2023Exploiting Programmatic Behavior of LLMs: Dual-Use Through Standard Security Attackscited by 278
- 2023Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Modelscited by 273
- 2024Is DPO Superior to PPO for LLM Alignment? A Comprehensive Studycited by 262
- 2023A Wolf in Sheep's Clothing: Generalized Nested Jailbreak Prompts can Fool Large Language Models Easilycited by 259
- 2022Is Reinforcement Learning (Not) for Natural Language Processing: Benchmarks, Baselines, and Building Blocks for Natural Language Policy Optimizationcited by 246
- 2024ArtPrompt: ASCII Art-based Jailbreak Attacks against Aligned LLMscited by 216
- 2023Scalable and Transferable Black-Box Jailbreaks for Language Models via Persona Modulationcited by 214
- 2023On the Robustness of ChatGPT: An Adversarial and Out-of-distribution Perspectivecited by 211
- 2024Rainbow Teaming: Open-Ended Generation of Diverse Adversarial Promptscited by 205
- 2024AutoDAN-Turbo: A Lifelong Agent for Strategy Self-Exploration to Jailbreak LLMscited by 199
- 2023Automatically Auditing Large Language Models via Discrete Optimizationcited by 191
- 2023How Robust is Google's Bard to Adversarial Image Attacks?cited by 183
- 2024COLD-Attack: Jailbreaking LLMs with Stealthiness and Controllabilitycited by 175
- 2023Image Hijacks: Adversarial Images can Control Generative Models at Runtimecited by 172
- 2024Arithmetic Control of LLMs for Diverse User Preferences: Directional Preference Alignment with Multi-Objective Rewardscited by 148
- 2023On the Adversarial Robustness of Multi-Modal Foundation Modelscited by 145
- 2024Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Modelscited by 136
- 2024Jailbreaking Attack against Multimodal Large Language Modelcited by 135
- 2023Open Sesame! Universal Black Box Jailbreaking of Large Language Modelscited by 134
- 2024Attacks, Defenses and Evaluations for LLM Conversation Safety: A Surveycited by 130
- 2023Abusing Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMscited by 124
50 most cited of 79 All 79 (JSON)
Built on by
Not cited yet.