@inproceedings{zhu-etal-2024-clean,
title = "{CLEAN}{--}{EVAL}: Clean Evaluation on Contaminated Large Language Models",
author = "Zhu, Wenhong and
Hao, Hongkun and
He, Zhiwei and
Song, Yun-Ze and
Yueyang, Jiao and
Zhang, Yumeng and
Hu, Hanxu and
Wei, Yiran and
Wang, Rui and
Lu, Hongyuan",
editor = "Duh, Kevin and
Gomez, Helena and
Bethard, Steven",
booktitle = "Findings of the Association for Computational Linguistics: NAACL 2024",
month = jun,
year = "2024",
address = "Mexico City, Mexico",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-naacl.53",
doi = "10.18653/v1/2024.findings-naacl.53",
pages = "835--847",
abstract = "We are currently in an era of fierce competition among various large language models (LLMs), continuously pushing the boundaries of benchmark performance. However, genuinely assessing the capabilities of these LLMs has become a challenging and critical issue due to potential data contamination. In this paper, we propose a novel and valuable method, Clean-Eval, which mitigates the issue of data contamination and evaluates the LLMs more cleanly. Clean-Eval employs a neural-based model to paraphrase and back-translate the contaminated data into a candidate set, generating expressions with the same meaning but in different surface forms. A semantic detector is then used to filter those generated low-quality samples to narrow down this candidate set. Candidates with moderate BLEURT scores against the original samples are selected as the final evaluation set. According to human assessment, this set is almost semantically equivalent to the original contamination set but expressed differently. We conduct experiments on 20 existing benchmarks across diverse tasks, and results demonstrate that Clean-Eval substantially restores the actual evaluation results on contaminated LLMs under both few-shot learning and fine-tuning scenarios.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhu-etal-2024-clean">
<titleInfo>
<title>CLEAN–EVAL: Clean Evaluation on Contaminated Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wenhong</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hongkun</namePart>
<namePart type="family">Hao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhiwei</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yun-Ze</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiao</namePart>
<namePart type="family">Yueyang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yumeng</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hanxu</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yiran</namePart>
<namePart type="family">Wei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rui</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hongyuan</namePart>
<namePart type="family">Lu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-06</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: NAACL 2024</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kevin</namePart>
<namePart type="family">Duh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Helena</namePart>
<namePart type="family">Gomez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steven</namePart>
<namePart type="family">Bethard</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Mexico City, Mexico</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>We are currently in an era of fierce competition among various large language models (LLMs), continuously pushing the boundaries of benchmark performance. However, genuinely assessing the capabilities of these LLMs has become a challenging and critical issue due to potential data contamination. In this paper, we propose a novel and valuable method, Clean-Eval, which mitigates the issue of data contamination and evaluates the LLMs more cleanly. Clean-Eval employs a neural-based model to paraphrase and back-translate the contaminated data into a candidate set, generating expressions with the same meaning but in different surface forms. A semantic detector is then used to filter those generated low-quality samples to narrow down this candidate set. Candidates with moderate BLEURT scores against the original samples are selected as the final evaluation set. According to human assessment, this set is almost semantically equivalent to the original contamination set but expressed differently. We conduct experiments on 20 existing benchmarks across diverse tasks, and results demonstrate that Clean-Eval substantially restores the actual evaluation results on contaminated LLMs under both few-shot learning and fine-tuning scenarios.</abstract>
<identifier type="citekey">zhu-etal-2024-clean</identifier>
<identifier type="doi">10.18653/v1/2024.findings-naacl.53</identifier>
<location>
<url>https://aclanthology.org/2024.findings-naacl.53</url>
</location>
<part>
<date>2024-06</date>
<extent unit="page">
<start>835</start>
<end>847</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CLEAN–EVAL: Clean Evaluation on Contaminated Large Language Models
%A Zhu, Wenhong
%A Hao, Hongkun
%A He, Zhiwei
%A Song, Yun-Ze
%A Yueyang, Jiao
%A Zhang, Yumeng
%A Hu, Hanxu
%A Wei, Yiran
%A Wang, Rui
%A Lu, Hongyuan
%Y Duh, Kevin
%Y Gomez, Helena
%Y Bethard, Steven
%S Findings of the Association for Computational Linguistics: NAACL 2024
%D 2024
%8 June
%I Association for Computational Linguistics
%C Mexico City, Mexico
%F zhu-etal-2024-clean
%X We are currently in an era of fierce competition among various large language models (LLMs), continuously pushing the boundaries of benchmark performance. However, genuinely assessing the capabilities of these LLMs has become a challenging and critical issue due to potential data contamination. In this paper, we propose a novel and valuable method, Clean-Eval, which mitigates the issue of data contamination and evaluates the LLMs more cleanly. Clean-Eval employs a neural-based model to paraphrase and back-translate the contaminated data into a candidate set, generating expressions with the same meaning but in different surface forms. A semantic detector is then used to filter those generated low-quality samples to narrow down this candidate set. Candidates with moderate BLEURT scores against the original samples are selected as the final evaluation set. According to human assessment, this set is almost semantically equivalent to the original contamination set but expressed differently. We conduct experiments on 20 existing benchmarks across diverse tasks, and results demonstrate that Clean-Eval substantially restores the actual evaluation results on contaminated LLMs under both few-shot learning and fine-tuning scenarios.
%R 10.18653/v1/2024.findings-naacl.53
%U https://aclanthology.org/2024.findings-naacl.53
%U https://doi.org/10.18653/v1/2024.findings-naacl.53
%P 835-847
Markdown (Informal)
[CLEAN–EVAL: Clean Evaluation on Contaminated Large Language Models](https://aclanthology.org/2024.findings-naacl.53) (Zhu et al., Findings 2024)
ACL
- Wenhong Zhu, Hongkun Hao, Zhiwei He, Yun-Ze Song, Jiao Yueyang, Yumeng Zhang, Hanxu Hu, Yiran Wei, Rui Wang, and Hongyuan Lu. 2024. CLEAN–EVAL: Clean Evaluation on Contaminated Large Language Models. In Findings of the Association for Computational Linguistics: NAACL 2024, pages 835–847, Mexico City, Mexico. Association for Computational Linguistics.