CHOMPS – Second Workshop on Confabulation, Hallucinations, & Overgeneration in Multilingual & Precision-critical Scenarios (with our apologies for cross-posting)
Venue: EACL 2027 (https://2027.eacl.org/), Athens, Greece Date: 9-14th March 2027 (TBC) Workshop website: https://chomps-workshop.github.io/ * Description * Large Language Models (LLMs) are becoming increasingly integrated into our everyday lives, but their tendency to produce hallucinations remains a major obstacle to their reliable use. LLMs frequently generate confident-sounding falsehoods that can be costly to verify, while their probabilistic nature makes preventing ungrounded outputs particularly challenging. These behaviors undermine trust, especially in precision-critical domains and under-resourced languages. Developing solutions that account for the diverse contexts in which LLMs are deployed is therefore essential to building more reliable and inclusive language technologies. The second iteration of the CHOMPS workshop will continue fostering discussion on hallucinations, confabulations, and overgeneration in LLMs, bringing together researchers and practitioners to explore their causes, consequences, and potential mitigation strategies. This year, we particularly welcome work on hallucination detection, output controllability and the mitigation of overgeneration, and factual alignment in multilingual and high-risk applications, with a special emphasis on challenges in the medical domain. Our goal is to bridge empirical research and real-world needs, advancing language technologies that are both trustworthy and fit for purpose. * Workshop topics * We invite submissions on, but not limited to, the following topics: - Metrics, benchmarks, and tools for hallucination detection - Factuality challenges in precision-critical and domain-specific applications, particularly medical, legal, and biotech settings - Hallucination mitigation strategies during inference and model training - Output controllability and approaches to reducing overgeneration - Hallucinatory and confabulatory behaviors in cross-lingual and multilingual settings - Factual alignment and reliability in under-resourced languages - Hallucinations in language and multimodal models (vision, text, and speech) - Medical-domain hallucinations, their evaluation, and their real-world consequences - Perspectives and case studies from other disciplines * Invited speakers * - Isabelle Augenstein, University of Copenhagen, Denmark - Iacer Calixto, University of Amsterdam, Netherlands * Submission details * The workshop is designed with a widely inclusive submission policy so as to foster as vibrant a discussion as possible. - Archival or non-archival submissions may consist of up to 8 pages (long) or 4 pages (short) of content. - Dissemination submissions may consist of up to 1 page of content. On acceptance, authors may add one additional page to accommodate changes suggested by the reviewers. Please use the ACL style templates available here: https://github.com/acl-org/acl-style-files The submissions need to be done in PDF format via: (a) via Direct submission (TBA) (b) via ARR commitment (TBA) * Shared Task: SHROOM-Visions 2.0 * Building on the success of SHROOM-Visions at UncertaiNLP 2026<https://helsinki-nlp.github.io/shroom/2026>, SHROOM-Visions 2.0 is coming to EACL 2027, co-located with CHOMPS! This edition poses a key question: Can your hallucination detector generalize to entirely unseen languages and domains? More details will follow soon. To participate, please register here: https://forms.gle/h2BfwdCFawDNCMNQ6 Visit the official CHOMPS website<https://chomps-workshop.github.io/> for more updates! * Important dates * First call for papers: October 13, 2026 Second call for papers: November 13, 2026 Paper submission deadline: December 15, 2026 Direct ARR commitment: December 22, 2026 Author notification: January 5, 2027 Camera-Ready due: January 19, 2027 Workshop date: March 9-14, 2027 (TBC) * Contact * For questions, please send an email to [email protected]<mailto:[email protected]> or contact one of the workshop chairs: - Aman Sinha, Université de Genève, [email protected]<mailto:[email protected]> - Chuyuan Li, University of Grenoble-Alps, [email protected]<mailto:[email protected]> - Raúl Vázquez, University of Helsinki, [email protected]<mailto:[email protected]> - Boya Zhang, Université de Genève, [email protected]<mailto:[email protected]>
_______________________________________________ Corpora mailing list -- [email protected] https://list.elra.info/mailman3/postorius/lists/corpora.list.elra.info/ To unsubscribe send an email to [email protected]
