Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMs
Identifikátory výsledku
Kód výsledku v IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F24%3A10492848" target="_blank" >RIV/00216208:11320/24:10492848 - isvavai.cz</a>
Výsledek na webu
<a href="https://aclanthology.org/2024.eacl-long.5" target="_blank" >https://aclanthology.org/2024.eacl-long.5</a>
DOI - Digital Object Identifier
—
Alternativní jazyky
Jazyk výsledku
angličtina
Název v původním jazyce
Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMs
Popis výsledku v původním jazyce
Natural Language Processing (NLP) research is increasingly focusing on the use of Large Language Models (LLMs), with some of the most popular ones being either fully or partially closed-source. The lack of access to model details, especially regarding training data, has repeatedly raised concerns about data contamination among researchers. Several attempts have been made to address this issue, but they are limited to anecdotal evidence and trial and error. Additionally, they overlook the problem of indirect data leaking, where models are iteratively improved by using data coming from users. In this work, we conduct the first systematic analysis of work using OpenAI's GPT-3.5 and GPT-4, the most prominently used LLMs today, in the context of data contamination. By analysing 255 papers and considering OpenAI's data usage policy, we extensively document the amount of data leaked to these models during the fi rst year after the model's release. We report that these models have been globally exposed to TILDE OPERATOR+D914.
Název v anglickém jazyce
Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMs
Popis výsledku anglicky
Natural Language Processing (NLP) research is increasingly focusing on the use of Large Language Models (LLMs), with some of the most popular ones being either fully or partially closed-source. The lack of access to model details, especially regarding training data, has repeatedly raised concerns about data contamination among researchers. Several attempts have been made to address this issue, but they are limited to anecdotal evidence and trial and error. Additionally, they overlook the problem of indirect data leaking, where models are iteratively improved by using data coming from users. In this work, we conduct the first systematic analysis of work using OpenAI's GPT-3.5 and GPT-4, the most prominently used LLMs today, in the context of data contamination. By analysing 255 papers and considering OpenAI's data usage policy, we extensively document the amount of data leaked to these models during the fi rst year after the model's release. We report that these models have been globally exposed to TILDE OPERATOR+D914.
Klasifikace
Druh
D - Stať ve sborníku
CEP obor
—
OECD FORD obor
10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)
Návaznosti výsledku
Projekt
—
Návaznosti
R - Projekt Ramcoveho programu EK
Ostatní
Rok uplatnění
2024
Kód důvěrnosti údajů
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Údaje specifické pro druh výsledku
Název statě ve sborníku
Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics
ISBN
979-8-89176-088-2
ISSN
—
e-ISSN
—
Počet stran výsledku
27
Strana od-do
67-93
Název nakladatele
Association for Computational Linguistics
Místo vydání
Kerrville, TX, USA
Místo konání akce
St. Julians, Malta
Datum konání akce
17. 3. 2024
Typ akce podle státní příslušnosti
WRD - Celosvětová akce
Kód UT WoS článku
—