@inproceedings{ren-etal-2024-learn,
title = "{I} Learn Better If You Speak My Language: Understanding the Superior Performance of Fine-Tuning Large Language Models with {LLM}-Generated Responses",
author = "Ren, Xuan and
Wu, Biao and
Liu, Lingqiao",
editor = "Al-Onaizan, Yaser and
Bansal, Mohit and
Chen, Yun-Nung",
booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2024",
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.emnlp-main.571/",
doi = "10.18653/v1/2024.emnlp-main.571",
pages = "10225--10245",
abstract = "This paper explores an intriguing observation: fine-tuning a large language model (LLM) with responses generated by a LLM often yields better results than using responses generated by humans, particularly in reasoning tasks. We conduct an in-depth investigation to understand why this occurs. Contrary to the common belief that these instances is due to the more detailed nature of LLM-generated content, our study identifies another contributing factor: an LLM is inherently more ``familiar'' with LLM generated responses. This familiarity is evidenced by lower perplexity before fine-tuning. We design a series of experiments to understand the impact of the ``familiarity'' and our conclusion reveals that this ``familiarity'' significantly impacts learning performance. Training with LLM-generated responses not only enhances performance but also helps maintain the model{'}s capabilities in other reasoning tasks after fine-tuning on a specific task."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ren-etal-2024-learn">
<titleInfo>
<title>I Learn Better If You Speak My Language: Understanding the Superior Performance of Fine-Tuning Large Language Models with LLM-Generated Responses</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xuan</namePart>
<namePart type="family">Ren</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Biao</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lingqiao</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yaser</namePart>
<namePart type="family">Al-Onaizan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohit</namePart>
<namePart type="family">Bansal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yun-Nung</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Miami, Florida, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>This paper explores an intriguing observation: fine-tuning a large language model (LLM) with responses generated by a LLM often yields better results than using responses generated by humans, particularly in reasoning tasks. We conduct an in-depth investigation to understand why this occurs. Contrary to the common belief that these instances is due to the more detailed nature of LLM-generated content, our study identifies another contributing factor: an LLM is inherently more “familiar” with LLM generated responses. This familiarity is evidenced by lower perplexity before fine-tuning. We design a series of experiments to understand the impact of the “familiarity” and our conclusion reveals that this “familiarity” significantly impacts learning performance. Training with LLM-generated responses not only enhances performance but also helps maintain the model’s capabilities in other reasoning tasks after fine-tuning on a specific task.</abstract>
<identifier type="citekey">ren-etal-2024-learn</identifier>
<identifier type="doi">10.18653/v1/2024.emnlp-main.571</identifier>
<location>
<url>https://aclanthology.org/2024.emnlp-main.571/</url>
</location>
<part>
<date>2024-11</date>
<extent unit="page">
<start>10225</start>
<end>10245</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T I Learn Better If You Speak My Language: Understanding the Superior Performance of Fine-Tuning Large Language Models with LLM-Generated Responses
%A Ren, Xuan
%A Wu, Biao
%A Liu, Lingqiao
%Y Al-Onaizan, Yaser
%Y Bansal, Mohit
%Y Chen, Yun-Nung
%S Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
%D 2024
%8 November
%I Association for Computational Linguistics
%C Miami, Florida, USA
%F ren-etal-2024-learn
%X This paper explores an intriguing observation: fine-tuning a large language model (LLM) with responses generated by a LLM often yields better results than using responses generated by humans, particularly in reasoning tasks. We conduct an in-depth investigation to understand why this occurs. Contrary to the common belief that these instances is due to the more detailed nature of LLM-generated content, our study identifies another contributing factor: an LLM is inherently more “familiar” with LLM generated responses. This familiarity is evidenced by lower perplexity before fine-tuning. We design a series of experiments to understand the impact of the “familiarity” and our conclusion reveals that this “familiarity” significantly impacts learning performance. Training with LLM-generated responses not only enhances performance but also helps maintain the model’s capabilities in other reasoning tasks after fine-tuning on a specific task.
%R 10.18653/v1/2024.emnlp-main.571
%U https://aclanthology.org/2024.emnlp-main.571/
%U https://doi.org/10.18653/v1/2024.emnlp-main.571
%P 10225-10245
Markdown (Informal)
[I Learn Better If You Speak My Language: Understanding the Superior Performance of Fine-Tuning Large Language Models with LLM-Generated Responses](https://aclanthology.org/2024.emnlp-main.571/) (Ren et al., EMNLP 2024)
ACL