Measuring and Reducing LLM Hallucination without Gold-Standard Answers
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866914827990990848 |
|---|---|
| author | Wei, Jiaheng Yao, Yuanshun Ton, Jean-Francois Guo, Hongyi Estornell, Andrew Liu, Yang |
| author_facet | Wei, Jiaheng Yao, Yuanshun Ton, Jean-Francois Guo, Hongyi Estornell, Andrew Liu, Yang |
| contents | LLM hallucination, i.e. generating factually incorrect yet seemingly convincing answers, is currently a major threat to the trustworthiness and reliability of LLMs. The first step towards solving this complicated problem is to measure it. However, existing hallucination metrics require having a benchmark dataset with gold-standard answers, i.e. "best" or "correct" answers written by humans. Such requirements make hallucination measurement costly and prone to human errors. In this work, we propose Factualness Evaluations via Weighting LLMs (FEWL), an innovative hallucination metric that is specifically designed for the scenario when gold-standard answers are absent. FEWL leverages the answers from off-the-shelf LLMs that serve as a proxy of gold-standard answers. The key challenge is how to quantify the expertise of reference LLMs resourcefully. We show FEWL has certain theoretical guarantees and demonstrate empirically it gives more accurate hallucination measures than naively using reference LLMs. We also show how to leverage FEWL to reduce hallucination through both in-context learning and supervised fine-tuning. Extensive experiment results on Truthful-QA, CHALE, and HaluEval datasets demonstrate the effectiveness of FEWL. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_10412 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Measuring and Reducing LLM Hallucination without Gold-Standard Answers Wei, Jiaheng Yao, Yuanshun Ton, Jean-Francois Guo, Hongyi Estornell, Andrew Liu, Yang Computation and Language Artificial Intelligence Machine Learning LLM hallucination, i.e. generating factually incorrect yet seemingly convincing answers, is currently a major threat to the trustworthiness and reliability of LLMs. The first step towards solving this complicated problem is to measure it. However, existing hallucination metrics require having a benchmark dataset with gold-standard answers, i.e. "best" or "correct" answers written by humans. Such requirements make hallucination measurement costly and prone to human errors. In this work, we propose Factualness Evaluations via Weighting LLMs (FEWL), an innovative hallucination metric that is specifically designed for the scenario when gold-standard answers are absent. FEWL leverages the answers from off-the-shelf LLMs that serve as a proxy of gold-standard answers. The key challenge is how to quantify the expertise of reference LLMs resourcefully. We show FEWL has certain theoretical guarantees and demonstrate empirically it gives more accurate hallucination measures than naively using reference LLMs. We also show how to leverage FEWL to reduce hallucination through both in-context learning and supervised fine-tuning. Extensive experiment results on Truthful-QA, CHALE, and HaluEval datasets demonstrate the effectiveness of FEWL. |
| title | Measuring and Reducing LLM Hallucination without Gold-Standard Answers |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2402.10412 |