FacLens: Transferable Probe for Foreseeing Non-Factuality in Fact-Seeking Question Answering of Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Yanling, Li, Haoyang, Zou, Hao, Zhang, Jing, He, Xinlei, Li, Qi, Xu, Ke
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918125755170816
author Wang, Yanling
Li, Haoyang
Zou, Hao
Zhang, Jing
He, Xinlei
Li, Qi
Xu, Ke
author_facet Wang, Yanling
Li, Haoyang
Zou, Hao
Zhang, Jing
He, Xinlei
Li, Qi
Xu, Ke
contents Despite advancements in large language models (LLMs), non-factual responses still persist in fact-seeking question answering. Unlike extensive studies on post-hoc detection of these responses, this work studies non-factuality prediction (NFP), predicting whether an LLM will generate a non-factual response prior to the response generation. Previous NFP methods have shown LLMs' awareness of their knowledge, but they face challenges in terms of efficiency and transferability. In this work, we propose a lightweight model named Factuality Lens (FacLens), which effectively probes hidden representations of fact-seeking questions for the NFP task. Moreover, we discover that hidden question representations sourced from different LLMs exhibit similar NFP patterns, enabling the transferability of FacLens across different LLMs to reduce development costs. Extensive experiments highlight FacLens's superiority in both effectiveness and efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2406_05328
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FacLens: Transferable Probe for Foreseeing Non-Factuality in Fact-Seeking Question Answering of Large Language Models
Wang, Yanling
Li, Haoyang
Zou, Hao
Zhang, Jing
He, Xinlei
Li, Qi
Xu, Ke
Computation and Language
Machine Learning
Despite advancements in large language models (LLMs), non-factual responses still persist in fact-seeking question answering. Unlike extensive studies on post-hoc detection of these responses, this work studies non-factuality prediction (NFP), predicting whether an LLM will generate a non-factual response prior to the response generation. Previous NFP methods have shown LLMs' awareness of their knowledge, but they face challenges in terms of efficiency and transferability. In this work, we propose a lightweight model named Factuality Lens (FacLens), which effectively probes hidden representations of fact-seeking questions for the NFP task. Moreover, we discover that hidden question representations sourced from different LLMs exhibit similar NFP patterns, enabling the transferability of FacLens across different LLMs to reduce development costs. Extensive experiments highlight FacLens's superiority in both effectiveness and efficiency.
title FacLens: Transferable Probe for Foreseeing Non-Factuality in Fact-Seeking Question Answering of Large Language Models
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2406.05328