@inproceedings{tefnik-kadlcik-2023-context,
title = "Can In-context Learners Learn a Reasoning Concept from Demonstrations?",
author = "{\v{S}}tef{\'a}nik, Michal and
Kadl{\v{c}}{\'\i}k, Marek",
editor = "Dalvi Mishra, Bhavana and
Durrett, Greg and
Jansen, Peter and
Neves Ribeiro, Danilo and
Wei, Jason",
booktitle = "Proceedings of the 1st Workshop on Natural Language Reasoning and Structured Explanations (NLRSE)",
month = jun,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.nlrse-1.8",
doi = "10.18653/v1/2023.nlrse-1.8",
pages = "107--115",
abstract = "Large language models show an emergent ability to learn a new task from a small number of input-output demonstrations. However, recent work shows that in-context learners largely rely on their pre-trained knowledge, such as the sentiment of the labels, instead of finding new associations in the input. However, the commonly-used few-shot evaluation settings using a random selection of in-context demonstrations can not disentangle models{'} ability to learn a new skill from demonstrations, as most of the randomly-selected demonstrations do not present relations informative for prediction beyond exposing the new task distribution. To disentangle models{'} in-context learning ability independent of models{'} memory, we introduce a Conceptual few-shot learning method selecting the demonstrations sharing a possibly-informative concept with the predicted sample. We extract a set of such concepts from annotated explanations and measure how much can models benefit from presenting these concepts in few-shot demonstrations. We find that smaller models are more sensitive to the presented concepts. While some of the models are able to benefit from concept-presenting demonstrations for each assessed concept, we find that none of the assessed in-context learners can benefit from all presented reasoning concepts consistently, leaving the in-context concept learning an open challenge.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="tefnik-kadlcik-2023-context">
<titleInfo>
<title>Can In-context Learners Learn a Reasoning Concept from Demonstrations?</title>
</titleInfo>
<name type="personal">
<namePart type="given">Michal</namePart>
<namePart type="family">Štefánik</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marek</namePart>
<namePart type="family">Kadlčík</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-06</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 1st Workshop on Natural Language Reasoning and Structured Explanations (NLRSE)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bhavana</namePart>
<namePart type="family">Dalvi Mishra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Greg</namePart>
<namePart type="family">Durrett</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peter</namePart>
<namePart type="family">Jansen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Danilo</namePart>
<namePart type="family">Neves Ribeiro</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jason</namePart>
<namePart type="family">Wei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Toronto, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Large language models show an emergent ability to learn a new task from a small number of input-output demonstrations. However, recent work shows that in-context learners largely rely on their pre-trained knowledge, such as the sentiment of the labels, instead of finding new associations in the input. However, the commonly-used few-shot evaluation settings using a random selection of in-context demonstrations can not disentangle models’ ability to learn a new skill from demonstrations, as most of the randomly-selected demonstrations do not present relations informative for prediction beyond exposing the new task distribution. To disentangle models’ in-context learning ability independent of models’ memory, we introduce a Conceptual few-shot learning method selecting the demonstrations sharing a possibly-informative concept with the predicted sample. We extract a set of such concepts from annotated explanations and measure how much can models benefit from presenting these concepts in few-shot demonstrations. We find that smaller models are more sensitive to the presented concepts. While some of the models are able to benefit from concept-presenting demonstrations for each assessed concept, we find that none of the assessed in-context learners can benefit from all presented reasoning concepts consistently, leaving the in-context concept learning an open challenge.</abstract>
<identifier type="citekey">tefnik-kadlcik-2023-context</identifier>
<identifier type="doi">10.18653/v1/2023.nlrse-1.8</identifier>
<location>
<url>https://aclanthology.org/2023.nlrse-1.8</url>
</location>
<part>
<date>2023-06</date>
<extent unit="page">
<start>107</start>
<end>115</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Can In-context Learners Learn a Reasoning Concept from Demonstrations?
%A Štefánik, Michal
%A Kadlčík, Marek
%Y Dalvi Mishra, Bhavana
%Y Durrett, Greg
%Y Jansen, Peter
%Y Neves Ribeiro, Danilo
%Y Wei, Jason
%S Proceedings of the 1st Workshop on Natural Language Reasoning and Structured Explanations (NLRSE)
%D 2023
%8 June
%I Association for Computational Linguistics
%C Toronto, Canada
%F tefnik-kadlcik-2023-context
%X Large language models show an emergent ability to learn a new task from a small number of input-output demonstrations. However, recent work shows that in-context learners largely rely on their pre-trained knowledge, such as the sentiment of the labels, instead of finding new associations in the input. However, the commonly-used few-shot evaluation settings using a random selection of in-context demonstrations can not disentangle models’ ability to learn a new skill from demonstrations, as most of the randomly-selected demonstrations do not present relations informative for prediction beyond exposing the new task distribution. To disentangle models’ in-context learning ability independent of models’ memory, we introduce a Conceptual few-shot learning method selecting the demonstrations sharing a possibly-informative concept with the predicted sample. We extract a set of such concepts from annotated explanations and measure how much can models benefit from presenting these concepts in few-shot demonstrations. We find that smaller models are more sensitive to the presented concepts. While some of the models are able to benefit from concept-presenting demonstrations for each assessed concept, we find that none of the assessed in-context learners can benefit from all presented reasoning concepts consistently, leaving the in-context concept learning an open challenge.
%R 10.18653/v1/2023.nlrse-1.8
%U https://aclanthology.org/2023.nlrse-1.8
%U https://doi.org/10.18653/v1/2023.nlrse-1.8
%P 107-115
Markdown (Informal)
[Can In-context Learners Learn a Reasoning Concept from Demonstrations?](https://aclanthology.org/2023.nlrse-1.8) (Štefánik & Kadlčík, NLRSE 2023)
ACL