2024年2月6日 · Knowledge hallucination have raised widespread concerns for the security and reliability of deployed LLMs. Previous efforts in detecting hallucinations have been employed at logit-level uncertainty estimation or language-level self-consistency evaluation, where the semantic information is inevitably lost during the token-decoding procedure.

Knowledge hallucination have raised widespread concerns for the security and reliability of deployed LLMs. Previous efforts in detecting hallucinations have been employed at logit-level uncertainty estimation or language-level self-consistency evaluation, where the semantic information is inevitably lost during the token-decoding procedure.

2025年2月15日 · 本研究提出了一个 inside 框架，利用 llm 内部状态中保留的语义信息进行幻觉检测。 具体来说，本文提出了一个既简单又有效的 EigenScore，用以衡量不同生成结果在嵌入空间中的语义一致性。

本文假设llm的内部状态可以用来揭示陈述的真实性。 因此，引入一种简单而有效的方法来检测LLM生成语句的真实性，该方法利用LLM的隐层激活来确定语句的真实性。

2024年7月11日 · While large language models (LLMs) have demonstrated remarkable abilities across various fields, hallucination remains a significant challenge. Recent studies have explored hallucinations through the lens of internal representations, proposing mechanisms to decipher LLMs' adherence to facts.

2024年2月6日 · Evidence that the LLM's internal state can be used to reveal the truthfulness of statements is provided, highlighting its potential to enhance the reliability of LLM-generated content and its practical applicability in real-world scenarios.

2024年7月3日 · Our empirical analysis reveals two key insights: (1) LLM internal states indicate whether they have seen the query in training data or not; and (2) LLM internal states show they are likely to hallucinate or not regarding the query.

5 天之前 · While large language models (LLMs) have demonstrated remarkable abilities across various fields, hallucination remains a significant challenge. Recent studies have explored hallucinations through the lens of internal representations, proposing mechanisms to decipher LLMs’ adherence to facts.

Thus, we propose to explore the dense semantic information retained within LLMs' INternal States for hallucInation DEtection (INSIDE). In particular, a simple yet effective EigenScore metric is proposed to better evaluate responses' self-consistency, which exploits the eigenvalues of responses' covariance matrix to measure the semantic ...

This work presents an INSIDE framework to exploit the semantic information that are retained within the internal states of LLMs for hallucination detection. Specifically, a simple yet effective EigenScore is proposed to measure the semantic consistency across different generations in the embedding space.