Think Twice: Measuring the Efficiency of Eliminating Prediction Shortcuts of Question Answering Models

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216224%3A14330%2F24%3A00135399" target="_blank" >RIV/00216224:14330/24:00135399 - isvavai.cz</a>

Výsledek na webu

<a href="https://aclanthology.org/2024.eacl-long.133.pdf" target="_blank" >https://aclanthology.org/2024.eacl-long.133.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Think Twice: Measuring the Efficiency of Eliminating Prediction Shortcuts of Question Answering Models

Popis výsledku v původním jazyce

While the Large Language Models (LLMs) dominate a majority of language understanding tasks, previous work shows that some of these results are supported by modelling spurious correlations of training datasets. Authors commonly assess model robustness by evaluating their models on out-of-distribution (OOD) datasets of the same task, but these datasets might share the bias of the training dataset. We propose a simple method for measuring a scale of models' reliance on any identified spurious feature and assess the robustness towards a large set of known and newly found prediction biases for various pre-trained models and debiasing methods in Question Answering (QA). We find that the reported OOD gains of debiasing methods can not be explained by mitigated reliance on biased features, suggesting that biases are shared among QA datasets. We further evidence this by measuring that performance of OOD models depends on bias features comparably to the ID model, motivating future work to refine the reports of LLMs' robustness to a level of known spurious features.

Název v anglickém jazyce

Think Twice: Measuring the Efficiency of Eliminating Prediction Shortcuts of Question Answering Models

Popis výsledku anglicky

While the Large Language Models (LLMs) dominate a majority of language understanding tasks, previous work shows that some of these results are supported by modelling spurious correlations of training datasets. Authors commonly assess model robustness by evaluating their models on out-of-distribution (OOD) datasets of the same task, but these datasets might share the bias of the training dataset. We propose a simple method for measuring a scale of models' reliance on any identified spurious feature and assess the robustness towards a large set of known and newly found prediction biases for various pre-trained models and debiasing methods in Question Answering (QA). We find that the reported OOD gains of debiasing methods can not be explained by mitigated reliance on biased features, suggesting that biases are shared among QA datasets. We further evidence this by measuring that performance of OOD models depends on bias features comparably to the ID model, motivating future work to refine the reports of LLMs' robustness to a level of known spurious features.

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)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2024

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)

ISBN

9798891760882

ISSN

—

e-ISSN

—

Počet stran výsledku

15

Strana od-do

2179-2193

Název nakladatele

Association for Computational Linguistics

Místo vydání

St. Julian's, Malta

Místo konání akce

St. Julian's, Malta

Datum konání akce

17. 3. 2024

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

001356732602016