Kernel Language Entropy: Fine-grained Uncertainty Quantification for LLMs from Semantic Similarities

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Nikitin, Alexander, Kossen, Jannik, Gal, Yarin, Marttinen, Pekka
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916266191618048
author Nikitin, Alexander
Kossen, Jannik
Gal, Yarin
Marttinen, Pekka
author_facet Nikitin, Alexander
Kossen, Jannik
Gal, Yarin
Marttinen, Pekka
contents Uncertainty quantification in Large Language Models (LLMs) is crucial for applications where safety and reliability are important. In particular, uncertainty can be used to improve the trustworthiness of LLMs by detecting factually incorrect model responses, commonly called hallucinations. Critically, one should seek to capture the model's semantic uncertainty, i.e., the uncertainty over the meanings of LLM outputs, rather than uncertainty over lexical or syntactic variations that do not affect answer correctness. To address this problem, we propose Kernel Language Entropy (KLE), a novel method for uncertainty estimation in white- and black-box LLMs. KLE defines positive semidefinite unit trace kernels to encode the semantic similarities of LLM outputs and quantifies uncertainty using the von Neumann entropy. It considers pairwise semantic dependencies between answers (or semantic clusters), providing more fine-grained uncertainty estimates than previous methods based on hard clustering of answers. We theoretically prove that KLE generalizes the previous state-of-the-art method called semantic entropy and empirically demonstrate that it improves uncertainty quantification performance across multiple natural language generation datasets and LLM architectures.
format Preprint
id arxiv_https___arxiv_org_abs_2405_20003
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Kernel Language Entropy: Fine-grained Uncertainty Quantification for LLMs from Semantic Similarities
Nikitin, Alexander
Kossen, Jannik
Gal, Yarin
Marttinen, Pekka
Machine Learning
Artificial Intelligence
Computation and Language
Uncertainty quantification in Large Language Models (LLMs) is crucial for applications where safety and reliability are important. In particular, uncertainty can be used to improve the trustworthiness of LLMs by detecting factually incorrect model responses, commonly called hallucinations. Critically, one should seek to capture the model's semantic uncertainty, i.e., the uncertainty over the meanings of LLM outputs, rather than uncertainty over lexical or syntactic variations that do not affect answer correctness. To address this problem, we propose Kernel Language Entropy (KLE), a novel method for uncertainty estimation in white- and black-box LLMs. KLE defines positive semidefinite unit trace kernels to encode the semantic similarities of LLM outputs and quantifies uncertainty using the von Neumann entropy. It considers pairwise semantic dependencies between answers (or semantic clusters), providing more fine-grained uncertainty estimates than previous methods based on hard clustering of answers. We theoretically prove that KLE generalizes the previous state-of-the-art method called semantic entropy and empirically demonstrate that it improves uncertainty quantification performance across multiple natural language generation datasets and LLM architectures.
title Kernel Language Entropy: Fine-grained Uncertainty Quantification for LLMs from Semantic Similarities
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2405.20003