SubmissionNumber#=%=#256
FinalPaperTitle#=%=#MALTO at SemEval-2024 Task 6: Leveraging Synthetic Data for LLM Hallucination Detection
ShortPaperTitle#=%=#
NumberOfPages#=%=#7
CopyrightSigned#=%=#Federico Borra
JobTitle#==#
Organization#==#Politecnico di Torino
Abstract#==#In Natural Language Generation (NLG), contemporary Large Language Models (LLMs) face several challenges, such as generating fluent yet inaccurate outputs and reliance on fluency-centric metrics. 
This often leads to neural networks exhibiting "hallucinations.'' 
The SHROOM challenge focuses on automatically identifying these hallucinations in the generated text. 
To tackle these issues, we introduce two key components, a data augmentation pipeline incorporating LLM-assisted pseudo-labelling and sentence rephrasing, and a voting ensemble from three models pre-trained on Natural Language Inference (NLI) tasks and fine-tuned on diverse datasets.
Author{1}{Firstname}#=%=#Federico
Author{1}{Lastname}#=%=#Borra
Author{1}{Username}#=%=#nemoomen
Author{1}{Email}#=%=#rico.borra@gmail.com
Author{1}{Affiliation}#=%=#Politecnico di Torino
Author{2}{Firstname}#=%=#Claudio
Author{2}{Lastname}#=%=#Savelli
Author{2}{Email}#=%=#claudio.savelli@studenti.polito.it
Author{2}{Affiliation}#=%=#Politecnico di Torino
Author{3}{Firstname}#=%=#Giacomo
Author{3}{Lastname}#=%=#Rosso
Author{3}{Email}#=%=#s309273@studenti.polito.it
Author{3}{Affiliation}#=%=#Politecnico di Torino
Author{4}{Firstname}#=%=#Alkis
Author{4}{Lastname}#=%=#Koudounas
Author{4}{Email}#=%=#alkis.koudounas@polito.it
Author{4}{Affiliation}#=%=#Politecnico di Torino
Author{5}{Firstname}#=%=#Flavio
Author{5}{Lastname}#=%=#Giobergia
Author{5}{Email}#=%=#flavio.giobergia@polito.it
Author{5}{Affiliation}#=%=#Politecnico di Torino

==========
èéáğö