Skip to main navigation Skip to search Skip to main content

Compos Mentis at SemEval2024 Task6: A Multi-Faceted Role-based Large Language Model Ensemble to Detect Hallucination

  • SUNY Buffalo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations

Abstract

Hallucinations in large language models (LLMs), where they generate fluent but factually incorrect outputs, pose challenges for applications requiring strict truthfulness. This work proposes a multi-faceted approach to detect such hallucinations across various language tasks. We leverage automatic data annotation using a proprietary LLM, fine-tuning of the Mistral-7B-instruct-v0.2 model on annotated and benchmark data, role-based and rationale-based prompting strategies, and an ensemble method combining different model outputs through majority voting. This comprehensive framework aims to improve the robustness and reliability of hallucination detection for LLM generations. Code and data.

Original languageEnglish
Title of host publicationSemEval 2024 - 18th International Workshop on Semantic Evaluation, Proceedings of the Workshop
EditorsAtul Kr. Ojha, A. Seza Dohruoz, Harish Tayyar Madabushi, Giovanni Da San Martino, Sara Rosenthal, Aiala Rosa
PublisherAssociation for Computational Linguistics (ACL)
Pages1449-1454
Number of pages6
ISBN (Electronic)9798891761070
DOIs
StatePublished - 2024
Event18th International Workshop on Semantic Evaluation, SemEval 2024, co-located with the 2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics, NAACL 2024 - Hybrid, Mexico City, Mexico
Duration: Jun 20 2024Jun 21 2024

Publication series

NameSemEval 2024 - 18th International Workshop on Semantic Evaluation, Proceedings of the Workshop

Conference

Conference18th International Workshop on Semantic Evaluation, SemEval 2024, co-located with the 2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics, NAACL 2024
Country/TerritoryMexico
CityHybrid, Mexico City
Period06/20/2406/21/24

Fingerprint

Dive into the research topics of 'Compos Mentis at SemEval2024 Task6: A Multi-Faceted Role-based Large Language Model Ensemble to Detect Hallucination'. Together they form a unique fingerprint.

Cite this