2024 HallucinationDiversityAwareActi

• (Xia et al., 2024) ⇒ Yu Xia, Xu Liu, Tong Yu, Sungchul Kim, Ryan A. Rossi, Anup Rao, Tung Mai, and Shuai Li. (2024). “Hallucination Diversity-Aware Active Learning for Text Summarization.” doi:10.48550/arXiv.2404.01588

Subject Headings: LLM Hallucination.

Notes

• The article investigates the potential of ChatGPT in extracting norms such as commitments, prohibitions, authorizations, and powers from contracts, aiming to understand the governance of interactions within multi-agent systems.
• The article demonstrates ChatGPT's ability to perform norm extraction effectively without the need for specific training or fine-tuning on legal documents, a significant advantage given the scarcity of annotated legal data.
• The article identifies key limitations in ChatGPT's performance, including the overlooking of crucial details, hallucination of information not present, incorrect parsing of conjunctions, and inaccuracies in norm elements or types.
• The article emphasizes the importance of crafting detailed and precise prompts for guiding ChatGPT's extraction process, highlighting the role of human input in achieving more accurate results.
• The article addresses the challenge of limited availability of high-quality, annotated datasets for contract understanding, pointing out the necessity for diverse and expertly annotated datasets to advance this field of study.
• The article suggests avenues for future research, including improving the accuracy of norm extraction, addressing the identified limitations of ChatGPT, and examining the impact of data quality on the model's performance.
• The article contributes to the discourse on applying artificial intelligence in legal studies by showcasing the potential and current challenges of using large language models like ChatGPT for legal document analysis.

Cited By

• http://scholar.google.com/scholar?q=%222024%22+Hallucination+Diversity-Aware+Active+Learning+for+Text+Summarization

Quotes

Abstract

Large Language Models (LLMs) have shown propensity to generate hallucinated outputs, i.e., texts that are factually incorrect or unsupported. Existing methods for alleviating hallucinations typically require costly human annotations to identify and correct hallucinations in LLM outputs. Moreover, most of these methods focus on a specific type of hallucination, e.g., entity or token errors, which limits their effectiveness in addressing various types of hallucinations exhibited in LLM outputs. To our best knowledge, in this paper we propose the first active learning framework to alleviate LLM hallucinations, reducing costly human annotations of hallucination needed. By measuring fine-grained hallucinations from errors in semantic frame, discourse and content verifiability in text summarization, we propose HAllucination Diversity-Aware Sampling (HADAS) to select diverse hallucinations for annotations in active learning for LLM finetuning. Extensive experiments on three datasets and different backbone models demonstrate advantages of our method in effectively and [[efficiently mitigating LLM hallucinations.

References

;

	Author	volume	Date Value	title	type	journal	titleUrl	doi	note	year
2024 HallucinationDiversityAwareActi	Yu Xia Xu Liu Tong Yu Sungchul Kim Ryan A. Rossi Anup Rao Tung Mai Shuai Li			Hallucination Diversity-Aware Active Learning for Text Summarization				10.48550/arXiv.2404.01588		2024