Computer Science
SemEval-2024 Task 6: SHROOM, a Shared-task on Hallucinations and Related Observable Overgeneration Mistakes
T. Mickus, E. Zosa, et al.
SHROOM presents a shared task on detecting hallucinations—fluent but inaccurate NLG outputs—using a newly constructed dataset of 4,000 model outputs labeled by five annotators across machine translation, paraphrase generation, and definition modeling. The research was conducted by Timothee Mickus, Elaine Zosa, Raúl Vázquez, Teemu Vahtola, Jörg Tiedemann, Vincent Segonne, Alessandro Raganato, and Marianna Apidianaki.
Related Publications
Explore these studies to deepen your understanding of the subject.

