@inproceedings{chen-etal-2023-relation,
title = "On the Relation between Sensitivity and Accuracy in In-Context Learning",
author = "Chen, Yanda and
Zhao, Chen and
Yu, Zhou and
McKeown, Kathleen and
He, He",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.12",
doi = "10.18653/v1/2023.findings-emnlp.12",
pages = "155--167",
abstract = "In-context learning (ICL) suffers from oversensitivity to the prompt, making it unreliable in real-world scenarios. We study the sensitivity of ICL with respect to multiple perturbation types. First, we find that label bias obscures the true sensitivity, and therefore prior work may have significantly underestimated ICL sensitivity. Second, we observe a strong negative correlation between ICL sensitivity and accuracy: predictions sensitive to perturbations are less likely to be correct. Motivated by these findings, we propose $SenSel$, a few-shot selective prediction method that abstains from sensitive predictions. Experiments on ten classification datasets show that $SenSel$ consistently outperforms two commonly used confidence-based and entropy-based baselines on abstention decisions.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2023-relation">
<titleInfo>
<title>On the Relation between Sensitivity and Accuracy in In-Context Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yanda</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chen</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhou</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kathleen</namePart>
<namePart type="family">McKeown</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">He</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Juan</namePart>
<namePart type="family">Pino</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kalika</namePart>
<namePart type="family">Bali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Singapore</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>In-context learning (ICL) suffers from oversensitivity to the prompt, making it unreliable in real-world scenarios. We study the sensitivity of ICL with respect to multiple perturbation types. First, we find that label bias obscures the true sensitivity, and therefore prior work may have significantly underestimated ICL sensitivity. Second, we observe a strong negative correlation between ICL sensitivity and accuracy: predictions sensitive to perturbations are less likely to be correct. Motivated by these findings, we propose SenSel, a few-shot selective prediction method that abstains from sensitive predictions. Experiments on ten classification datasets show that SenSel consistently outperforms two commonly used confidence-based and entropy-based baselines on abstention decisions.</abstract>
<identifier type="citekey">chen-etal-2023-relation</identifier>
<identifier type="doi">10.18653/v1/2023.findings-emnlp.12</identifier>
<location>
<url>https://aclanthology.org/2023.findings-emnlp.12</url>
</location>
<part>
<date>2023-12</date>
<extent unit="page">
<start>155</start>
<end>167</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T On the Relation between Sensitivity and Accuracy in In-Context Learning
%A Chen, Yanda
%A Zhao, Chen
%A Yu, Zhou
%A McKeown, Kathleen
%A He, He
%Y Bouamor, Houda
%Y Pino, Juan
%Y Bali, Kalika
%S Findings of the Association for Computational Linguistics: EMNLP 2023
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F chen-etal-2023-relation
%X In-context learning (ICL) suffers from oversensitivity to the prompt, making it unreliable in real-world scenarios. We study the sensitivity of ICL with respect to multiple perturbation types. First, we find that label bias obscures the true sensitivity, and therefore prior work may have significantly underestimated ICL sensitivity. Second, we observe a strong negative correlation between ICL sensitivity and accuracy: predictions sensitive to perturbations are less likely to be correct. Motivated by these findings, we propose SenSel, a few-shot selective prediction method that abstains from sensitive predictions. Experiments on ten classification datasets show that SenSel consistently outperforms two commonly used confidence-based and entropy-based baselines on abstention decisions.
%R 10.18653/v1/2023.findings-emnlp.12
%U https://aclanthology.org/2023.findings-emnlp.12
%U https://doi.org/10.18653/v1/2023.findings-emnlp.12
%P 155-167
Markdown (Informal)
[On the Relation between Sensitivity and Accuracy in In-Context Learning](https://aclanthology.org/2023.findings-emnlp.12) (Chen et al., Findings 2023)
ACL