ETF: An Entity Tracing Framework for Hallucination Detection in Code Summaries

Gespeichert in:

Bibliographische Detailangaben
Veröffentlicht in:	arXiv.org (Dec 18, 2024), p. n/a
1. Verfasser:	Maharaj, Kishan
Weitere Verfasser:	Munigala, Vitobha, Tamilselvam, Srikanth G, Kumar, Prince, Sen, Sayandeep, Kodeswaran, Palani, Mishra, Abhijit, Bhattacharyya, Pushpak
Veröffentlicht:	Cornell University Library, arXiv.org
Schlagworte:	Program verification (computers) Large language models Summaries Tracing Task complexity
Online-Zugang:	Citation/Abstract Full text outside of ProQuest
Tags:	Tag hinzufügen Keine Tags, Fügen Sie das erste Tag hinzu!

Beschreibung
Abstract:	Recent advancements in large language models (LLMs) have significantly enhanced their ability to understand both natural language and code, driving their use in tasks like natural language-to-code (NL2Code) and code summarization. However, LLMs are prone to hallucination-outputs that stray from intended meanings. Detecting hallucinations in code summarization is especially difficult due to the complex interplay between programming and natural languages. We introduce a first-of-its-kind dataset with \(\sim\)10K samples, curated specifically for hallucination detection in code summarization. We further propose a novel Entity Tracing Framework (ETF) that a) utilizes static program analysis to identify code entities from the program and b) uses LLMs to map and verify these entities and their intents within generated code summaries. Our experimental analysis demonstrates the effectiveness of the framework, leading to a 0.73 F1 score. This approach provides an interpretable method for detecting hallucinations by grounding entities, allowing us to evaluate summary accuracy.
ISSN:	2331-8422
Quelle:	Engineering Database