@inproceedings{chandler-etal-2025-deloitte,
    title = "Deloitte (Drocks) at {S}em{E}val-2025 Task 3: Fine-Grained Multi-lingual Hallucination Detection Using Internal {LLM} Weights",
    author = "Chandler, Alex  and
      Abburi, Harika  and
      Bhattacharya, Sanmitra  and
      Bowen, Edward  and
      Pudota, Nirmala",
    editor = "Rosenthal, Sara  and
      Ros{\'a}, Aiala  and
      Ghosh, Debanjan  and
      Zampieri, Marcos",
    booktitle = "Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025)",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/transition-to-people-yaml/2025.semeval-1.144/",
    pages = "1089--1097",
    ISBN = "979-8-89176-273-2",
    abstract = "Large Language Models (LLMs) have greatly advanced the field of Natural Language Generation (NLG). Despite their remarkable capabilities, their tendency to hallucinate{---}producing inaccurate or misleading information-remains a barrier to wider adoption. Current hallucination detection methods mainly employ coarse-grained binary classification at the sentence or document level, overlooking the need for precise identification of the specific text spans containing hallucinations. In this paper, we proposed a methodology that generates supplementary context and processes text using an LLM to extract internal weights (features) from various layers. These extracted features serve as input for a neural network classifier designed to perform token-level binary detection of hallucinations. Subsequently, we map the resulting token-level predictions to character-level predictions, enabling the identification of spans of hallucinated text, which we refer to as hallucination spans. Our model achieved a top-ten ranking in 13 of the 14 languages and secured first place for the French language in the SemEval: Multilingual Shared-task on Hallucinations and Related Observable Overgeneration Mistakes (Mu-SHROOM), utilizing the Mu-SHROOM dataset provided by the task organizers."
}