{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,17]],"date-time":"2026-04-17T04:23:25Z","timestamp":1776399805131,"version":"3.51.2"},"reference-count":171,"publisher":"MIT Press","license":[{"start":{"date-parts":[[2025,7,2]],"date-time":"2025-07-02T00:00:00Z","timestamp":1751414400000},"content-version":"vor","delay-in-days":182,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["direct.mit.edu"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2025,6,27]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Abstention, the refusal of large language models (LLMs) to provide an answer, is increasingly recognized for its potential to mitigate hallucinations and enhance safety in LLM systems. In this survey, we introduce a framework to examine abstention from three perspectives: the query, the model, and human values. We organize the literature on abstention methods, benchmarks, and evaluation metrics using this framework, and discuss merits and limitations of prior work. We further identify and motivate areas for future research, such as whether abstention can be achieved as a meta-capability that transcends specific tasks or domains, and opportunities to optimize abstention abilities in specific contexts. In doing so, we aim to broaden the scope and impact of abstention methodologies in AI systems.1<\/jats:p>","DOI":"10.1162\/tacl_a_00754","type":"journal-article","created":{"date-parts":[[2025,7,2]],"date-time":"2025-07-02T20:07:03Z","timestamp":1751486823000},"page":"529-556","update-policy":"https:\/\/doi.org\/10.1162\/mitpressjournals.corrections.policy","source":"Crossref","is-referenced-by-count":14,"title":["Know Your Limits: A Survey of Abstention in Large Language Models"],"prefix":"10.1162","volume":"13","author":[{"given":"Bingbing","family":"Wen","sequence":"first","affiliation":[{"name":"University of Washington, USA. bingbw@cs.washington.edu"}]},{"given":"Jihan","family":"Yao","sequence":"additional","affiliation":[{"name":"University of Washington, USA. jihany2@cs.washington.edu"}]},{"given":"Shangbin","family":"Feng","sequence":"additional","affiliation":[{"name":"University of Washington, USA. shangbin@cs.washington.edu"}]},{"given":"Chenjun","family":"Xu","sequence":"additional","affiliation":[{"name":"University of Washington, USA. chenjux@cs.washington.edu"}]},{"given":"Yulia","family":"Tsvetkov","sequence":"additional","affiliation":[{"name":"University of Washington, USA. yuliats@cs.washington.edu"}]},{"given":"Bill","family":"Howe","sequence":"additional","affiliation":[{"name":"University of Washington, USA. billhowe@cs.washington.edu"}]},{"given":"Lucy Lu","family":"Wang","sequence":"additional","affiliation":[{"name":"University of Washington, USA. lucylw@cs.washington.edu"},{"name":"Allen Institute for AI, USA"}]}],"member":"281","published-online":{"date-parts":[[2025,6,27]]},"reference":[{"key":"2025070216065737100_bib1","article-title":"GPT-4 technical report","author":"Achiam","year":"2023","journal-title":"arXiv preprint arXiv:2303.08774"},{"key":"2025070216065737100_bib2","article-title":"Distinguishing the knowable from the unknowable with language models","author":"Ahdritz","year":"2024","journal-title":"arXiv preprint arXiv:2402.03563"},{"key":"2025070216065737100_bib3","doi-asserted-by":"publisher","first-page":"475","DOI":"10.1145\/3331184.3331265","article-title":"Asking clarifying questions in open-domain information-seeking conversations","volume-title":"Proceedings of the 42nd International ACM Sigir Conference on Research and Development in Information Retrieval","author":"Aliannejadi","year":"2019"},{"key":"2025070216065737100_bib4","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.383","article-title":"Knowledge of knowledge: Exploring known-unknowns uncertainty with large language models","author":"Amayuelas","year":"2023","journal-title":"arXiv preprint arXiv:2305.13712"},{"key":"2025070216065737100_bib5","article-title":"Introducing claude","author":"Anthropic","year":"2023"},{"key":"2025070216065737100_bib6","article-title":"Foundational challenges in assuring alignment and safety of large language models","author":"Anwar","year":"2024","journal-title":"arXiv preprint arXiv:2404.09932"},{"key":"2025070216065737100_bib7","doi-asserted-by":"publisher","first-page":"1492","DOI":"10.18653\/v1\/2021.acl-long.118","article-title":"Challenges in information-seeking QA: Unanswerable questions and paragraph retrieval","volume-title":"Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)","author":"Asai","year":"2021"},{"key":"2025070216065737100_bib8","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.findings-emnlp.68","article-title":"The internal state of an LLM knows when it\u2019s lying","volume-title":"The 2023 Conference on Empirical Methods in Natural Language Processing","author":"Azaria","year":"2023"},{"key":"2025070216065737100_bib9","article-title":"Training a helpful and harmless assistant with reinforcement learning from human feedback","author":"Bai","year":"2022"},{"key":"2025070216065737100_bib10","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.762","article-title":"Language models are Homer Simpson! Safety re-alignment of fine-tuned language models through task arithmetic","author":"Bhardwaj","year":"2024"},{"key":"2025070216065737100_bib11","article-title":"Safety-tuned LLaMAs: Lessons from improving the safety of large language models that follow instructions","volume-title":"The Twelfth International Conference on Learning Representations","author":"Bianchi","year":"2024"},{"key":"2025070216065737100_bib12","article-title":"The art of saying no: Contextual noncompliance in language models","author":"Brahman","year":"2024","journal-title":"arXiv preprint arXiv:2407.12043"},{"key":"2025070216065737100_bib13","article-title":"Defending against alignment-breaking attacks via robustly aligned LLM","author":"Cao","year":"2023","journal-title":"arXiv preprint arXiv:2309.14348"},{"key":"2025070216065737100_bib14","doi-asserted-by":"publisher","first-page":"3628","DOI":"10.18653\/v1\/2024.emnlp-main.212","article-title":"Learn to refuse: Making large language models more controllable and reliable through knowledge scope limitation and refusal mechanism","volume-title":"Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing","author":"Cao","year":"2024"},{"key":"2025070216065737100_bib15","article-title":"Jailbreakbench: An open robustness benchmark for jailbreaking large language models","author":"Chao","year":"2024","journal-title":"arXiv preprint arXiv:2404.01318"},{"key":"2025070216065737100_bib16","article-title":"Jailbreaking black box large language models in twenty queries","author":"Chao","year":"2023","journal-title":"arXiv preprint arXiv:2310.08419"},{"key":"2025070216065737100_bib17","doi-asserted-by":"publisher","first-page":"29","DOI":"10.1145\/3605760.3623764","article-title":"Jailbreaker in jail: Moving target defense for large language models","volume-title":"Proceedings of the 10th ACM Workshop on Moving Target Defense","author":"Chen","year":"2023"},{"key":"2025070216065737100_bib18","article-title":"INSIDE: LLMs\u2019 internal states retain the power of hallucination detection","volume-title":"The Twelfth International Conference on Learning Representations","author":"Chen","year":"2024"},{"key":"2025070216065737100_bib19","doi-asserted-by":"publisher","first-page":"5190","DOI":"10.18653\/v1\/2023.findings-emnlp.345","article-title":"Adaptation with self-evaluation to improve selective prediction in LLMs","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2023","author":"Chen","year":"2023"},{"key":"2025070216065737100_bib20","article-title":"Can AI assistants know what they don\u2019t know?","volume-title":"Forty-first International Conference on Machine Learning","author":"Cheng","year":"2024"},{"key":"2025070216065737100_bib21","article-title":"Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality","author":"Chiang","year":"2023"},{"key":"2025070216065737100_bib22","doi-asserted-by":"publisher","first-page":"2174","DOI":"10.18653\/v1\/D18-1241","article-title":"QuAC: Question answering in context","volume-title":"Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing","author":"Choi","year":"2018"},{"key":"2025070216065737100_bib23","article-title":"Palm: Scaling language modeling with pathways","author":"Chowdhery","year":"2022","journal-title":"arXiv preprint arXiv:2204.02311"},{"key":"2025070216065737100_bib24","doi-asserted-by":"publisher","first-page":"454","DOI":"10.1162\/tacl_a_00317","article-title":"TyDi QA: A benchmark for information- seeking question answering in typologically diverse languages","volume":"8","author":"Clark","year":"2020","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2025070216065737100_bib25","article-title":"Selectively answering ambiguous questions","author":"Cole","year":"2023","journal-title":"arXiv preprint arXiv:2305.14613"},{"key":"2025070216065737100_bib26","article-title":"Safe RLHF: Safe reinforcement learning from human feedback","volume-title":"The Twelfth International Conference on Learning Representations","author":"Dai","year":"2024"},{"key":"2025070216065737100_bib27","doi-asserted-by":"publisher","first-page":"4599","DOI":"10.18653\/v1\/2021.naacl-main.365","article-title":"A dataset of information-seeking questions and answers anchored in research papers","volume-title":"Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","author":"Dasigi","year":"2021"},{"key":"2025070216065737100_bib28","doi-asserted-by":"publisher","first-page":"13652","DOI":"10.18653\/v1\/2024.emnlp-main.757","article-title":"Don\u2019t just say \u201cI don\u2019t know\u201d! self-aligning large language models for responding to unknown questions with explanations","volume-title":"Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing","author":"Deng","year":"2024"},{"key":"2025070216065737100_bib29","doi-asserted-by":"publisher","first-page":"295","DOI":"10.18653\/v1\/2020.emnlp-main.21","article-title":"Calibration of pre-trained transformers","volume-title":"Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)","author":"Desai","year":"2020"},{"key":"2025070216065737100_bib30","first-page":"10088","article-title":"Qlora: Efficient finetuning of quantized llms","volume-title":"Advances in Neural Information Processing Systems","author":"Dettmers","year":"2023"},{"key":"2025070216065737100_bib31","doi-asserted-by":"publisher","first-page":"4537","DOI":"10.18653\/v1\/D19-1461","article-title":"Build it break it fix it for dialogue safety: Robustness from adversarial human attack","volume-title":"Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)","author":"Dinan","year":"2019"},{"key":"2025070216065737100_bib32","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.276","article-title":"Shifting attention to relevance: Towards the uncertainty estimation of large language models","author":"Duan","year":"2023","journal-title":"arXiv preprint arXiv:2307.01379"},{"key":"2025070216065737100_bib33","doi-asserted-by":"publisher","first-page":"345","DOI":"10.18653\/v1\/2021.emnlp-main.29","article-title":"Latent hatred: A benchmark for understanding implicit hate speech","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"ElSherief","year":"2021"},{"key":"2025070216065737100_bib34","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.239","article-title":"Teaching LLMs to abstain across languages via multilingual feedback","author":"Feng","year":"2024","journal-title":"arXiv preprint arXiv:2406.15948"},{"key":"2025070216065737100_bib35","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.786","article-title":"Don\u2019t hallucinate, abstain: Identifying LLM knowledge gaps via Multi-LLM collaboration","author":"Feng","year":"2024","journal-title":"arXiv preprint arXiv:2402.00367"},{"key":"2025070216065737100_bib36","first-page":"1050","article-title":"Dropout as a bayesian approximation: Representing model uncertainty in deep learning","volume-title":"Proceedings of The 33rd International Conference on Machine Learning","author":"Gal","year":"2016"},{"key":"2025070216065737100_bib37","first-page":"10835","article-title":"Scaling laws for reward model overoptimization","volume-title":"International Conference on Machine Learning","author":"Gao","year":"2023"},{"key":"2025070216065737100_bib38","doi-asserted-by":"publisher","first-page":"3356","DOI":"10.18653\/v1\/2020.findings-emnlp.301","article-title":"RealToxicityPrompts: Evaluating neural toxic degeneration in language models","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2020","author":"Gehman","year":"2020"},{"key":"2025070216065737100_bib39","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.444","article-title":"Does fine-tuning llms on new knowledge encourage hallucinations?","author":"Gekhman","year":"2024","journal-title":"arXiv preprint arXiv:2405.05904"},{"key":"2025070216065737100_bib40","first-page":"1321","article-title":"On calibration of modern neural networks","volume-title":"International Conference on Machine Learning","author":"Guo","year":"2017"},{"key":"2025070216065737100_bib41","first-page":"1321","article-title":"On calibration of modern neural networks","volume-title":"Proceedings of the 34th International Conference on Machine Learning","author":"Guo","year":"2017"},{"key":"2025070216065737100_bib42","article-title":"Controllable preference optimization: Toward controllable multi-objective alignment","author":"Guo","year":"2024","journal-title":"arXiv preprint arXiv:2402.19085"},{"key":"2025070216065737100_bib43","article-title":"Wildguard: Open one-stop moderation tools for safety risks, jailbreaks, and refusals of llms","author":"Han","year":"2024","journal-title":"arXiv preprint arXiv:2406.18495"},{"key":"2025070216065737100_bib44","doi-asserted-by":"publisher","first-page":"3309","DOI":"10.18653\/v1\/2022.acl-long.234","article-title":"ToxiGen: A large-scale machine-generated dataset for adversarial and implicit hate speech detection","volume-title":"Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Hartvigsen","year":"2022"},{"key":"2025070216065737100_bib45","article-title":"Deep learning scaling is predictable, empirically","author":"Hestness","year":"2017","journal-title":"arXiv preprint arXiv:1712.00409"},{"key":"2025070216065737100_bib46","article-title":"Training compute-optimal large language models","author":"Hoffmann","year":"2022","journal-title":"arXiv preprint arXiv: 2203.15556"},{"key":"2025070216065737100_bib47","article-title":"Decomposing uncertainty for large language models through input clarification ensembling","author":"Hou","year":"2023","journal-title":"arXiv preprint arXiv:2311.08718"},{"key":"2025070216065737100_bib48","article-title":"LoRA: Low-rank adaptation of large language models","volume-title":"International Conference on Learning Representations","author":"Edward","year":"2022"},{"key":"2025070216065737100_bib49","article-title":"Token-level adversarial prompt detection based on perplexity measures and contextual information","author":"Zhengmian","year":"2024"},{"key":"2025070216065737100_bib50","article-title":"Catastrophic jailbreak of open-source LLMs via exploiting generation","volume-title":"The Twelfth International Conference on Learning Representations","author":"Huang","year":"2024"},{"key":"2025070216065737100_bib51","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-emnlp.785","article-title":"Calibrating long-form generations from large language models","author":"Huang","year":"2024","journal-title":"arXiv preprint arXiv:2402.06544"},{"key":"2025070216065737100_bib52","article-title":"Baseline defenses for adversarial attacks against aligned language models","author":"Jain","year":"2023","journal-title":"arXiv preprint arXiv:2309.00614"},{"key":"2025070216065737100_bib53","article-title":"Defending large language models against jailbreak attacks via semantic smoothing","author":"Ji","year":"2024"},{"key":"2025070216065737100_bib54","article-title":"Beavertails: Towards improved safety alignment of LLM via a human-preference dataset","volume-title":"Thirty-seventh Conference on Neural Information Processing Systems Datasets and Benchmarks Track","author":"Ji","year":"2023"},{"issue":"12","key":"2025070216065737100_bib55","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3571730","article-title":"Survey of hallucination in natural language generation","volume":"55","author":"Ji","year":"2023","journal-title":"ACM Computing Surveys"},{"key":"2025070216065737100_bib56","doi-asserted-by":"publisher","first-page":"962","DOI":"10.1162\/tacl_a_00407","article-title":"How can we know when language models know? On the calibration of language models for question answering","volume":"9","author":"Jiang","year":"2021","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2025070216065737100_bib57","doi-asserted-by":"publisher","first-page":"2567","DOI":"10.18653\/v1\/D19-1259","article-title":"PubMedQA: A dataset for biomedical research question answering","volume-title":"Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)","author":"Jin","year":"2019"},{"key":"2025070216065737100_bib58","article-title":"Trust or escalate: Llm judges with provable guarantees for human agreement","author":"Jung","year":"2024","journal-title":"arXiv preprint arXiv:2407.18370"},{"key":"2025070216065737100_bib59","article-title":"Language models (mostly) know what they know","volume":"abs\/2207.05221","author":"Kadavath","year":"2022","journal-title":"ArXiv preprint"},{"key":"2025070216065737100_bib60","doi-asserted-by":"publisher","first-page":"5684","DOI":"10.18653\/v1\/2020.acl-main.503","article-title":"Selective question answering under domain shift","volume-title":"Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics","author":"Kamath","year":"2020"},{"key":"2025070216065737100_bib61","article-title":"Unfamiliar finetuning examples control how language models hallucinate","author":"Kang","year":"2024","journal-title":"arXiv preprint arXiv:2403.05612"},{"key":"2025070216065737100_bib62","article-title":"Scaling laws for neural language models","author":"Kaplan","year":"2020","journal-title":"arXiv preprint arXiv:2001.08361"},{"key":"2025070216065737100_bib63","article-title":"Realtime QA: What\u2019s the answer right now?","volume-title":"Thirty-seventh Conference on Neural Information Processing Systems Datasets and Benchmarks Track","author":"Kasai","year":"2023"},{"key":"2025070216065737100_bib64","article-title":"Break the breakout: Reinventing lm defense against jailbreak attacks with self-refinement","author":"Kim","year":"2024"},{"key":"2025070216065737100_bib65","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.751","article-title":"Epistemology of language models: Do language models have holistic knowledge?","author":"Kim","year":"2024","journal-title":"arXiv preprint arXiv:2403.12862"},{"key":"2025070216065737100_bib66","doi-asserted-by":"publisher","first-page":"822","DOI":"10.1145\/3630106.3658941","article-title":"\u201ci\u2019m not sure, but...\u201d: Examining the impact of large language models\u2019 uncertainty expression on user reliance and trust","volume-title":"Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency","author":"Kim","year":"2024"},{"key":"2025070216065737100_bib67","doi-asserted-by":"publisher","first-page":"2409","DOI":"10.18653\/v1\/2023.emnlp-main.148","article-title":"The past, present and better future of feedback learning in large language models for subjective human preferences and values","volume-title":"Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing","author":"Kirk","year":"2023"},{"key":"2025070216065737100_bib68","article-title":"Personalisation within bounds: A risk taxonomy and policy framework for the alignment of large language models with personalised feedback","author":"Kirk","year":"2023","journal-title":"arXiv preprint arXiv:2303.05453"},{"key":"2025070216065737100_bib69","article-title":"Semantic uncertainty: Linguistic invariances for uncertainty estimation in natural language generation","volume-title":"The Eleventh International Conference on Learning Representations","author":"Kuhn","year":"2023"},{"key":"2025070216065737100_bib70","article-title":"Certifying LLM safety against adversarial prompting","author":"Kumar","year":"2024"},{"key":"2025070216065737100_bib71","doi-asserted-by":"publisher","first-page":"452","DOI":"10.1162\/tacl_a_00276","article-title":"Natural questions: A benchmark for question answering research","volume":"7","author":"Kwiatkowski","year":"2019","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2025070216065737100_bib72","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.235","article-title":"Salad-bench: A hierarchical and comprehensive safety benchmark for large language models","author":"Li","year":"2024","journal-title":"arXiv preprint arXiv:2402.05044"},{"key":"2025070216065737100_bib73","article-title":"MediQ: Question-asking LLMs for adaptive and reliable medical reasoning","author":"Li","year":"2024","journal-title":"arXiv preprint arXiv:2406.00922"},{"key":"2025070216065737100_bib74","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.knowledgenlp-1.4","article-title":"Learning to trust your feelings: Leveraging self-awareness in LLMs for hallucination mitigation","author":"Liang","year":"2024"},{"key":"2025070216065737100_bib75","article-title":"Flame: Factuality-aware alignment for large language models","author":"Lin","year":"2024"},{"key":"2025070216065737100_bib76","article-title":"Teaching models to express their uncertainty in words","author":"Lin","year":"2022","journal-title":"arXiv preprint arXiv:2205.14334"},{"key":"2025070216065737100_bib77","article-title":"Generating with confidence: Uncertainty quantification for black-box large language models","author":"Lin","year":"2024"},{"key":"2025070216065737100_bib78","doi-asserted-by":"publisher","first-page":"4694","DOI":"10.18653\/v1\/2023.findings-emnlp.311","article-title":"ToxicChat: Unveiling hidden challenges of toxicity detection in real-world user-AI conversation","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2023","author":"Zi","year":"2023"},{"key":"2025070216065737100_bib79","article-title":"Keeping LLMs aligned after fine-tuning: The crucial role of prompt templates","volume-title":"ICLR 2024 Workshop on Reliable and Responsible Foundation Models","author":"Lyu","year":"2024"},{"key":"2025070216065737100_bib80","doi-asserted-by":"publisher","first-page":"9802","DOI":"10.18653\/v1\/2023.acl-long.546","article-title":"When not to trust language models: Investigating effectiveness of parametric and non-parametric memories","volume-title":"Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Mallen","year":"2023"},{"key":"2025070216065737100_bib81","article-title":"Harmbench: A standardized evaluation framework for automated red teaming and robust refusal","author":"Mazeika","year":"2024","journal-title":"arXiv preprint arXiv:2402.04249"},{"key":"2025070216065737100_bib82","doi-asserted-by":"publisher","first-page":"857","DOI":"10.1162\/tacl_a_00494","article-title":"Reducing conversational agents\u2019 overconfidence through linguistic calibration","volume":"10","author":"Mielke","year":"2022","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2025070216065737100_bib83","doi-asserted-by":"publisher","first-page":"5783","DOI":"10.18653\/v1\/2020.emnlp-main.466","article-title":"AmbigQA: Answering ambiguous open-domain questions","volume-title":"Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)","author":"Min","year":"2020"},{"key":"2025070216065737100_bib84","article-title":"Fight back against jailbreaking via prompt adversarial tuning","volume-title":"NeurIPS","author":"Mo","year":"2024"},{"key":"2025070216065737100_bib85","doi-asserted-by":"publisher","first-page":"10056","DOI":"10.18653\/v1\/2023.acl-long.559","article-title":"DisentQA: Disentangling parametric and contextual knowledge with counterfactual question answering","volume-title":"Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Neeman","year":"2023"},{"key":"2025070216065737100_bib86","unstructured":"OpenAI. 2023. Introducing ChatGPT."},{"key":"2025070216065737100_bib87","first-page":"27730","article-title":"Training language models to follow instructions with human feedback","volume-title":"Advances in Neural Information Processing Systems","author":"Ouyang","year":"2022"},{"key":"2025070216065737100_bib88","article-title":"LLM self defense: By self examination, LLMs know they are being tricked","volume-title":"The Second Tiny Papers Track at ICLR 2024","author":"Phute","year":"2024"},{"key":"2025070216065737100_bib89","article-title":"Bergeron: Combating adversarial attacks through a conscience-based alignment framework","author":"Pisano","year":"2023","journal-title":"arXiv preprint arXiv:2312.00029"},{"key":"2025070216065737100_bib90","doi-asserted-by":"publisher","first-page":"9558","DOI":"10.18653\/v1\/2021.emnlp-main.752","article-title":"ONION: A simple and effective defense against textual backdoor attacks","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"Qi","year":"2021"},{"key":"2025070216065737100_bib91","article-title":"Fine-tuning aligned language models compromises safety, even when users do not intend to!","volume-title":"The Twelfth International Conference on Learning Representations","author":"Qi","year":"2024"},{"key":"2025070216065737100_bib92","article-title":"Latent jailbreak: A benchmark for evaluating text safety and output robustness of large language models","author":"Qiu","year":"2023","journal-title":"arXiv preprint arXiv:2307.08487"},{"key":"2025070216065737100_bib93","article-title":"Direct preference optimization: Your language model is secretly a reward model","volume-title":"Thirty-seventh Conference on Neural Information Processing Systems","author":"Rafailov","year":"2023"},{"key":"2025070216065737100_bib94","doi-asserted-by":"publisher","first-page":"784","DOI":"10.18653\/v1\/P18-2124","article-title":"Know what you don\u2019t know: Unanswerable questions for SQuAD","volume-title":"Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)","author":"Rajpurkar","year":"2018"},{"key":"2025070216065737100_bib95","doi-asserted-by":"publisher","first-page":"249","DOI":"10.1162\/tacl_a_00266","article-title":"CoQA: A conversational question answering challenge","volume":"7","author":"Reddy","year":"2019","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2025070216065737100_bib96","article-title":"Gpqa: A graduate-level google-proof q&a benchmark","author":"Rein","year":"2023","journal-title":"arXiv preprint arXiv:2311.12022"},{"key":"2025070216065737100_bib97","first-page":"49","article-title":"Self-evaluation improves selective generation in large language models","volume-title":"Proceedings on \u201cI Can\u2019t Believe It\u2019s Not Better: Failure Modes in the Age of Foundation Models\u201d at NeurIPS 2023 Workshops","author":"Ren","year":"2023"},{"key":"2025070216065737100_bib98","article-title":"SmoothLLM: Defending large language models against jailbreaking attacks","author":"Robey","year":"2023"},{"key":"2025070216065737100_bib99","doi-asserted-by":"publisher","first-page":"5377","DOI":"10.18653\/v1\/2024.naacl-long.301","article-title":"XSTest: A test suite for identifying exaggerated safety behaviours in large language models","volume-title":"Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)","author":"R\u00f6ttger","year":"2024"},{"key":"2025070216065737100_bib100","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i26.34975","article-title":"Safetyprompts: A systematic review of open datasets for evaluating and improving large language model safety","author":"R\u00f6ttger","year":"2024","journal-title":"arXiv preprint arXiv:2404.05399"},{"key":"2025070216065737100_bib101","article-title":"Proximal policy optimization algorithms","author":"Schulman","year":"2017","journal-title":"arXiv preprint arXiv:1707.06347"},{"key":"2025070216065737100_bib102","doi-asserted-by":"publisher","first-page":"6138","DOI":"10.18653\/v1\/2021.emnlp-main.496","article-title":"Simple entity- centric questions challenge dense retrievers","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"Sciavolino","year":"2021"},{"key":"2025070216065737100_bib103","article-title":"Scalable and transferable black-box jailbreaks for language models via persona modulation","volume-title":"Socially Responsible Language Modelling Research","author":"Shah","year":"2023"},{"key":"2025070216065737100_bib104","doi-asserted-by":"publisher","first-page":"102433","DOI":"10.1016\/j.cose.2021.102433","article-title":"Bddr: An effective defense against textual backdoor attacks","volume":"110","author":"Shao","year":"2021","journal-title":"Computers & Security"},{"key":"2025070216065737100_bib105","doi-asserted-by":"publisher","DOI":"10.1145\/3658644.3670388","article-title":"\u201cdo anything now\u201d: Characterizing and evaluating in-the-wild jailbreak prompts on large language models","author":"Shen","year":"2023","journal-title":"arXiv preprint arXiv:2308.03825"},{"key":"2025070216065737100_bib106","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/2024.naacl-long.422","article-title":"Safer-instruct: Aligning language models with automated preference data","author":"Shi","year":"2024"},{"key":"2025070216065737100_bib107","article-title":"LLaMAs know what gpts don\u2019t show: Surrogate models for confidence estimation","author":"Shrivastava","year":"2023","journal-title":"arXiv preprint arXiv:2311.08877"},{"key":"2025070216065737100_bib108","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.findings-emnlp.552","article-title":"Getting moRE out of mixture of language model reasoning experts","volume-title":"The 2023 Conference on Empirical Methods in Natural Language Processing","author":"Si","year":"2023"},{"key":"2025070216065737100_bib109","doi-asserted-by":"publisher","first-page":"3607","DOI":"10.18653\/v1\/2023.emnlp-main.220","article-title":"The curious case of hallucinatory (un)answerability: Finding truths in the hidden states of over- confident large language models","volume-title":"Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing","author":"Slobodkin","year":"2023"},{"key":"2025070216065737100_bib110","article-title":"A strongreject for empty jailbreaks","author":"Souly","year":"2024","journal-title":"arXiv preprint arXiv:2402.10260"},{"key":"2025070216065737100_bib111","article-title":"Lacie: Listener-aware finetuning for confidence calibration in large language models","author":"Stengel-Eskin","year":"2024","journal-title":"arXiv preprint arXiv:2405.21028"},{"key":"2025070216065737100_bib112","doi-asserted-by":"publisher","first-page":"5257","DOI":"10.1609\/aaai.v37i4.25656","article-title":"Defending against backdoor attacks in natural language generation","volume-title":"Proceedings of the AAAI Conference on Artificial Intelligence","author":"Sun","year":"2023"},{"key":"2025070216065737100_bib113","article-title":"Salmon: Self-alignment with instructable reward models","author":"Sun","year":"2024"},{"key":"2025070216065737100_bib114","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.308","article-title":"Rethinking the inception architecture for computer vision","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Szegedy","year":"2016"},{"key":"2025070216065737100_bib115","doi-asserted-by":"publisher","first-page":"5433","DOI":"10.18653\/v1\/2023.emnlp-main.330","article-title":"Just ask for calibration: Strategies for eliciting calibrated confidence scores from language models fine-tuned with human feedback","volume-title":"Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing","author":"Tian","year":"2023"},{"key":"2025070216065737100_bib116","article-title":"Uncertainty-based abstention in LLMs improves safety and reduces hallucinations","author":"Tomani","year":"2024","journal-title":"arXiv preprint arXiv:2404.10960"},{"key":"2025070216065737100_bib117","article-title":"LLaMA: Open and efficient foundation language models","author":"Touvron","year":"2023","journal-title":"ArXiv preprint"},{"key":"2025070216065737100_bib118","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00475","article-title":"MuSiQue: Multihop questions via single-hop question composition","author":"Trivedi","year":"2022","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2025070216065737100_bib119","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.776","article-title":"The art of defending: A systematic evaluation and analysis of LLM defense strategies on safety and over- defensiveness","author":"Varshney","year":"2023","journal-title":"arXiv preprint arXiv:2401 .00287"},{"key":"2025070216065737100_bib120","doi-asserted-by":"publisher","first-page":"1995","DOI":"10.18653\/v1\/2022.findings-acl.158","article-title":"Investigating selective prediction approaches across several tasks in IID, OOD, and adversarial settings","volume-title":"Findings of the Association for Computational Linguistics: ACL 2022","author":"Varshney","year":"2022"},{"key":"2025070216065737100_bib121","article-title":"The instruction hierarchy: Training LLMs to prioritize privileged instructions","author":"Wallace","year":"2024","journal-title":"arXiv preprint arXiv:2404.13208"},{"key":"2025070216065737100_bib122","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.585","article-title":"Inferaligner: Inference-time alignment for harmlessness through cross-model guidance","author":"Wang","year":"2024","journal-title":"arXiv preprint arXiv:2401.11206"},{"key":"2025070216065737100_bib123","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.349","article-title":"All languages matter: On the multilingual safety of large language models","author":"Wang","year":"2023","journal-title":"arXiv preprint arXiv:2310.00905"},{"key":"2025070216065737100_bib124","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.948","article-title":"Defending LLMs against jailbreaking attacks via backtranslation","author":"Wang","year":"2024","journal-title":"arXiv preprint arXiv:2402.16459"},{"key":"2025070216065737100_bib125","article-title":"Resolving knowledge conflicts in large language models","volume":"abs\/2310.00935","author":"Wang","year":"2023","journal-title":"ArXiv"},{"key":"2025070216065737100_bib126","first-page":"896","article-title":"Do- not-answer: Evaluating safeguards in LLMs","volume-title":"Findings of the Association for Computational Linguistics: EACL 2024","author":"Wang","year":"2024"},{"key":"2025070216065737100_bib127","doi-asserted-by":"publisher","first-page":"1648","DOI":"10.18653\/v1\/2024.naacl-long.92","article-title":"SELF-GUARD: Empower the LLM to safeguard itself","volume-title":"Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)","author":"Wang","year":"2024"},{"key":"2025070216065737100_bib128","article-title":"Finetuned language models are zero-shot learners","volume-title":"International Conference on Learning Representations","author":"Wei","year":"2022"},{"key":"2025070216065737100_bib129","first-page":"24824","article-title":"Chain-of-thought prompting elicits reasoning in large language models","volume-title":"Advances in Neural Information Processing Systems","author":"Wei","year":"2022"},{"key":"2025070216065737100_bib130","article-title":"Jailbreak and guard aligned language models with only few in-context demonstrations","author":"Wei","year":"2024"},{"key":"2025070216065737100_bib131","doi-asserted-by":"publisher","first-page":"3437","DOI":"10.18653\/v1\/2024.findings-emnlp.197","article-title":"Characterizing LLM abstention behavior in science QA with context perturbations","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2024","author":"Wen","year":"2024"},{"key":"2025070216065737100_bib132","article-title":"Batchensemble: An alternative approach to efficient ensemble and lifelong learning","volume-title":"International Conference on Learning Representations","author":"Wen","year":"2020"},{"key":"2025070216065737100_bib133","doi-asserted-by":"publisher","DOI":"10.1145\/3613904.3642135","article-title":"\u201cas an AI language model, i cannot\u201d: Investigating llm denials of user requests","volume-title":"Proceedings of the CHI Conference on Human Factors in Computing Systems","author":"Wester","year":"2024"},{"key":"2025070216065737100_bib134","doi-asserted-by":"publisher","first-page":"148","DOI":"10.1007\/978-3-031-20059-5_9","article-title":"Reliable visual question answering: Abstain rather than answer incorrectly","volume-title":"European Conference on Computer Vision","author":"Whitehead","year":"2022"},{"key":"2025070216065737100_bib135","doi-asserted-by":"publisher","first-page":"1199","DOI":"10.1145\/3630106.3658966","article-title":"Laboratory-scale AI: Open-weight models are competitive with ChatGPT even in low-resource settings","volume-title":"Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency","author":"Wolfe","year":"2024"},{"key":"2025070216065737100_bib136","article-title":"Defending pre-trained language models as few-shot learners against backdoor attacks","volume-title":"Thirty-seventh Conference on Neural Information Processing Systems","author":"Xi","year":"2023"},{"key":"2025070216065737100_bib137","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.findings-emnlp.538","article-title":"Uncertainty quantification with pre-trained language models: A large-scale empirical analysis","author":"Xiao","year":"2022"},{"key":"2025070216065737100_bib138","article-title":"Sorry-bench: Systematically evaluating large language model safety refusal behaviors","author":"Xie","year":"2024","journal-title":"arXiv preprint arXiv:2406.14598"},{"issue":"12","key":"2025070216065737100_bib139","doi-asserted-by":"publisher","first-page":"1486","DOI":"10.1038\/s42256-023-00765-8","article-title":"Defending ChatGPT against jailbreak attack via self-reminders","volume":"5","author":"Xie","year":"2023","journal-title":"Nature Machine Intelligence"},{"key":"2025070216065737100_bib140","doi-asserted-by":"publisher","first-page":"1040","DOI":"10.18653\/v1\/2021.acl-long.84","article-title":"The art of abstention: Selective prediction and error regularization for natural language processing","volume-title":"Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)","author":"Ji","year":"2021"},{"key":"2025070216065737100_bib141","article-title":"Can LLMs express their uncertainty? An empirical evaluation of confidence elicitation in LLMs","volume-title":"The Twelfth International Conference on Learning Representations","author":"Xiong","year":"2024"},{"key":"2025070216065737100_bib142","first-page":"2390","article-title":"Detoxifying language models risks marginalizing minority voices","volume-title":"Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","author":"Albert","year":"2021"},{"key":"2025070216065737100_bib143","article-title":"Cvalues: Measuring the values of chinese large language models from safety to responsibility","author":"Guohai","year":"2023","journal-title":"arXiv preprint arXiv:2307.09705"},{"key":"2025070216065737100_bib144","article-title":"The earth is flat because...: Investigating LLMs\u2019 belief towards misinformation via persuasive conversation","author":"Rongwu","year":"2023","journal-title":"arXiv preprint arXiv:2312.09085"},{"key":"2025070216065737100_bib145","article-title":"Knowledge conflicts for llms: A survey","author":"Rongwu","year":"2024","journal-title":"arXiv preprint arXiv:2403.08319"},{"key":"2025070216065737100_bib146","article-title":"Alignment for honesty","author":"Yang","year":"2023","journal-title":"arXiv preprint arXiv:2312.07000"},{"key":"2025070216065737100_bib147","article-title":"A survey on recent advances in LLM-Based multi-turn dialogue systems","author":"Yi","year":"2024","journal-title":"arXiv preprint arXiv:2402.18013"},{"key":"2025070216065737100_bib148","doi-asserted-by":"publisher","first-page":"8653","DOI":"10.18653\/v1\/2023.findings-acl.551","article-title":"Do large language models know what they don\u2019t know?","volume-title":"Findings of the Association for Computational Linguistics: ACL 2023","author":"Yin","year":"2023"},{"key":"2025070216065737100_bib149","article-title":"Low-resource languages jailbreak gpt-4","author":"Yong","year":"2023","journal-title":"arXiv preprint arXiv:2310.02446"},{"key":"2025070216065737100_bib150","doi-asserted-by":"publisher","first-page":"1972","DOI":"10.18653\/v1\/2023.findings-eacl.150","article-title":"Selective-lama: Selective prediction for confidence-aware evaluation of language models","volume-title":"Findings of the Association for Computational Linguistics: EACL 2023","author":"Yoshikawa","year":"2023"},{"key":"2025070216065737100_bib151","article-title":"GPT-4 is too smart to be safe: Stealthy chat with LLMs via cipher","volume-title":"The Twelfth International Conference on Learning Representations","author":"Yuan","year":"2024"},{"key":"2025070216065737100_bib152","article-title":"RigorLLM: Resilient guardrails for large language models against undesired content","volume-title":"Forty-first International Conference on Machine Learning","author":"Yuan","year":"2024"},{"key":"2025070216065737100_bib153","doi-asserted-by":"publisher","first-page":"2980","DOI":"10.18653\/v1\/2023.findings-emnlp.197","article-title":"On uncertainty calibration and selective generation in probabilistic neural summarization: A benchmark study","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2023","author":"Zablotskaia","year":"2023"},{"key":"2025070216065737100_bib154","article-title":"Autodefense: Multi-agent LLM defense against jailbreak attacks","author":"Zeng","year":"2024","journal-title":"arXiv preprint arXiv:2403.04783"},{"key":"2025070216065737100_bib155","doi-asserted-by":"publisher","first-page":"7113","DOI":"10.18653\/v1\/2024.naacl-long.394","article-title":"R-tuning: Instructing large language models to say \u2018I don\u2019t know\u2019","volume-title":"Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)","author":"Zhang","year":"2024"},{"key":"2025070216065737100_bib156","article-title":"Controllable safety alignment: Inference-time adaptation to diverse safety requirements","author":"Zhang","year":"2024","journal-title":"arXiv preprint arXiv:2410.08968"},{"key":"2025070216065737100_bib157","doi-asserted-by":"publisher","first-page":"7371","DOI":"10.18653\/v1\/2021.emnlp-main.586","article-title":"SituatedQA: Incorporating extra-linguistic contexts into QA","volume-title":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","author":"Zhang","year":"2021"},{"key":"2025070216065737100_bib158","doi-asserted-by":"publisher","first-page":"39","DOI":"10.1162\/tacl_a_00632","article-title":"Benchmarking large language models for news summarization","volume":"12","author":"Zhang","year":"2023","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2025070216065737100_bib159","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.107","article-title":"Self-alignment for factuality: Mitigating hallucinations in LLMs via self-evaluation","author":"Zhang","year":"2024"},{"key":"2025070216065737100_bib160","article-title":"Intention analysis makes LLMs a good jailbreak defender","author":"Zhang","year":"2024"},{"key":"2025070216065737100_bib161","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.481","article-title":"Defending large language models against jailbreaking attacks through goal prioritization","author":"Zhang","year":"2023","journal-title":"arXiv preprint arXiv:2311.09096"},{"key":"2025070216065737100_bib162","doi-asserted-by":"publisher","first-page":"4207","DOI":"10.18653\/v1\/2024.emnlp-main.242","article-title":"I could\u2019ve asked that: Reformulating unanswerable questions","volume-title":"Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing","author":"Zhao","year":"2024"},{"key":"2025070216065737100_bib163","article-title":"Calibrating sequence likelihood improves conditional language generation","volume-title":"The Eleventh International Conference on Learning Representations","author":"Zhao","year":"2022"},{"key":"2025070216065737100_bib164","doi-asserted-by":"publisher","first-page":"7051","DOI":"10.18653\/v1\/2024.naacl-long.390","article-title":"Knowing what LLMs DO NOT know: A simple yet effective self- detection method","volume-title":"Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)","author":"Zhao","year":"2024"},{"key":"2025070216065737100_bib165","article-title":"Prompt-driven llm safeguarding via directed representation optimization","author":"Zheng","year":"2024","journal-title":"arXiv preprint arXiv:2401.18018"},{"key":"2025070216065737100_bib166","article-title":"Robust prompt optimization for defending language models against jailbreaking attacks","author":"Zhou","year":"2024"},{"key":"2025070216065737100_bib167","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.198","article-title":"Relying on the unreliable: The impact of language models\u2019 reluctance to express uncertainty","author":"Zhou","year":"2024","journal-title":"arXiv preprint arXiv:2401.06730"},{"key":"2025070216065737100_bib168","doi-asserted-by":"publisher","first-page":"5506","DOI":"10.18653\/v1\/2023.emnlp-main.335","article-title":"Navigating the grey area: How expressions of uncertainty and overconfidence affect language models","volume-title":"Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing","author":"Zhou","year":"2023"},{"key":"2025070216065737100_bib169","doi-asserted-by":"publisher","first-page":"14544","DOI":"10.18653\/v1\/2023.findings-emnlp.968","article-title":"Context-faithful prompting for large language models","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2023","author":"Zhou","year":"2023"},{"key":"2025070216065737100_bib170","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.1121","article-title":"Defending jailbreak prompts via in-context adversarial game","author":"Zhou","year":"2024","journal-title":"arXiv preprint arXiv:2402 .13148"},{"key":"2025070216065737100_bib171","article-title":"Universal and transferable adversarial attacks on aligned language models","author":"Zou","year":"2023","journal-title":"arXiv preprint arXiv:2307.15043"}],"container-title":["Transactions of the Association for Computational Linguistics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/direct.mit.edu\/tacl\/article-pdf\/doi\/10.1162\/tacl_a_00754\/2534960\/tacl_a_00754.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/direct.mit.edu\/tacl\/article-pdf\/doi\/10.1162\/tacl_a_00754\/2534960\/tacl_a_00754.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,7,2]],"date-time":"2025-07-02T20:07:21Z","timestamp":1751486841000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/tacl\/article\/doi\/10.1162\/tacl_a_00754\/131566\/Know-Your-Limits-A-Survey-of-Abstention-in-Large"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025]]},"references-count":171,"URL":"https:\/\/doi.org\/10.1162\/tacl_a_00754","relation":{},"ISSN":["2307-387X"],"issn-type":[{"value":"2307-387X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2025]]},"published":{"date-parts":[[2025]]}}}