Enregistré dans:
Détails bibliographiques
Auteurs principaux: Huang, Junsheng, He, Zhitao, Huang, Yucheng, Polisetty, Sandeep, Wang, Qingyun, Fung, Yi. R
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2504.21773
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914034876416000
author Huang, Junsheng
He, Zhitao
Huang, Yucheng
Polisetty, Sandeep
Wang, Qingyun
Fung, Yi. R
author_facet Huang, Junsheng
He, Zhitao
Huang, Yucheng
Polisetty, Sandeep
Wang, Qingyun
Fung, Yi. R
contents The hallucination of non-existent facts by LLMs is an important problem given its widespread adoption across various applications. Previous research addresses this problem by analyzing the internal parameterized knowledge boundaries to estimate confidence. However, these studies focus on the single-problem setting and have not explored the more challenging multi-problem setting, which requires accurately answering multiple questions simultaneously. We introduce a novel method for the multi-problem setting, Multiple Answers and Confidence Stepwise Tuning (MAC-Tuning), that separates the learning of answer prediction and confidence estimation during fine-tuning on instruction data. Extensive experiments demonstrate that our method outperforms baselines by up to 25\% in average precision.
format Preprint
id arxiv_https___arxiv_org_abs_2504_21773
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MAC-Tuning: LLM Multi-Compositional Problem Reasoning with Enhanced Knowledge Boundary Awareness
Huang, Junsheng
He, Zhitao
Huang, Yucheng
Polisetty, Sandeep
Wang, Qingyun
Fung, Yi. R
Computation and Language
Artificial Intelligence
The hallucination of non-existent facts by LLMs is an important problem given its widespread adoption across various applications. Previous research addresses this problem by analyzing the internal parameterized knowledge boundaries to estimate confidence. However, these studies focus on the single-problem setting and have not explored the more challenging multi-problem setting, which requires accurately answering multiple questions simultaneously. We introduce a novel method for the multi-problem setting, Multiple Answers and Confidence Stepwise Tuning (MAC-Tuning), that separates the learning of answer prediction and confidence estimation during fine-tuning on instruction data. Extensive experiments demonstrate that our method outperforms baselines by up to 25\% in average precision.
title MAC-Tuning: LLM Multi-Compositional Problem Reasoning with Enhanced Knowledge Boundary Awareness
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2504.21773