MlingConf: A Comprehensive Study of Multilingual Confidence Estimation on Large Language Models

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Xue, Boyang, Wang, Hongru, Wang, Rui, Wang, Sheng, Wang, Zezhong, Du, Yiming, Liang, Bin, Wong, Kam-Fai
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910655835013120
author Xue, Boyang
Wang, Hongru
Wang, Rui
Wang, Sheng
Wang, Zezhong
Du, Yiming
Liang, Bin
Wong, Kam-Fai
author_facet Xue, Boyang
Wang, Hongru
Wang, Rui
Wang, Sheng
Wang, Zezhong
Du, Yiming
Liang, Bin
Wong, Kam-Fai
contents The tendency of Large Language Models (LLMs) to generate hallucinations raises concerns regarding their reliability. Therefore, confidence estimations indicating the extent of trustworthiness of the generations become essential. However, current LLM confidence estimations in languages other than English remain underexplored. This paper addresses this gap by introducing a comprehensive investigation of Multilingual Confidence estimation (MlingConf) on LLMs, focusing on both language-agnostic (LA) and language-specific (LS) tasks to explore the performance and language dominance effects of multilingual confidence estimations on different tasks. The benchmark comprises four meticulously checked and human-evaluate high-quality multilingual datasets for LA tasks and one for the LS task tailored to specific social, cultural, and geographical contexts of a language. Our experiments reveal that on LA tasks English exhibits notable linguistic dominance in confidence estimations than other languages, while on LS tasks, using question-related language to prompt LLMs demonstrates better linguistic dominance in multilingual confidence estimations. The phenomena inspire a simple yet effective native-tone prompting strategy by employing language-specific prompts for LS tasks, effectively improving LLMs' reliability and accuracy on LS tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2410_12478
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MlingConf: A Comprehensive Study of Multilingual Confidence Estimation on Large Language Models
Xue, Boyang
Wang, Hongru
Wang, Rui
Wang, Sheng
Wang, Zezhong
Du, Yiming
Liang, Bin
Wong, Kam-Fai
Computation and Language
The tendency of Large Language Models (LLMs) to generate hallucinations raises concerns regarding their reliability. Therefore, confidence estimations indicating the extent of trustworthiness of the generations become essential. However, current LLM confidence estimations in languages other than English remain underexplored. This paper addresses this gap by introducing a comprehensive investigation of Multilingual Confidence estimation (MlingConf) on LLMs, focusing on both language-agnostic (LA) and language-specific (LS) tasks to explore the performance and language dominance effects of multilingual confidence estimations on different tasks. The benchmark comprises four meticulously checked and human-evaluate high-quality multilingual datasets for LA tasks and one for the LS task tailored to specific social, cultural, and geographical contexts of a language. Our experiments reveal that on LA tasks English exhibits notable linguistic dominance in confidence estimations than other languages, while on LS tasks, using question-related language to prompt LLMs demonstrates better linguistic dominance in multilingual confidence estimations. The phenomena inspire a simple yet effective native-tone prompting strategy by employing language-specific prompts for LS tasks, effectively improving LLMs' reliability and accuracy on LS tasks.
title MlingConf: A Comprehensive Study of Multilingual Confidence Estimation on Large Language Models
topic Computation and Language
url https://arxiv.org/abs/2410.12478