<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liu-etal-2026-notes">
    <titleInfo>
        <title>Notes to Self: Can LLMs Benefit from Experiential Abstractions?</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Chang</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Xinyu</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Artur</namePart>
        <namePart type="family">Dubrawski</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-462-0</identifier>
    </relatedItem>
    <abstract>Humans distill experience into reusable abstractions, e.g., strategies and cautionary reminders, and apply them to gradually solve problems more effectively. We study whether Large Language Models (LLMs) can similarly benefit from such experiential abstractions. From LLMs’ solution traces on a training set, a stronger teacher or the LLMs themselves extract natural-language abstractions into a retrievable library. We explore two modes of abstraction usage: (1) inference-time retrieval and (2) reinforcement learning (RL) with abstraction-augmented training prompts. Experiential abstractions improve LLM performance on mathematical and logical reasoning benchmarks. Self-extracted abstractions match teacher-extracted ones, and our abstraction usage framework can transfer to other datasets and models. These findings suggest LLMs can extract and apply experiential abstractions much as humans leverage distilled experience.</abstract>
    <identifier type="citekey">liu-etal-2026-notes</identifier>
    <identifier type="doi">10.18653/2026.findings-emnlp.2050</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-emnlp.2050/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>43634</start>
            <end>43645</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Notes to Self: Can LLMs Benefit from Experiential Abstractions?
%A Liu, Chang
%A Li, Xinyu
%A Dubrawski, Artur
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F liu-etal-2026-notes
%X Humans distill experience into reusable abstractions, e.g., strategies and cautionary reminders, and apply them to gradually solve problems more effectively. We study whether Large Language Models (LLMs) can similarly benefit from such experiential abstractions. From LLMs’ solution traces on a training set, a stronger teacher or the LLMs themselves extract natural-language abstractions into a retrievable library. We explore two modes of abstraction usage: (1) inference-time retrieval and (2) reinforcement learning (RL) with abstraction-augmented training prompts. Experiential abstractions improve LLM performance on mathematical and logical reasoning benchmarks. Self-extracted abstractions match teacher-extracted ones, and our abstraction usage framework can transfer to other datasets and models. These findings suggest LLMs can extract and apply experiential abstractions much as humans leverage distilled experience.
%R 10.18653/2026.findings-emnlp.2050
%U https://aclanthology.org/2026.findings-emnlp.2050/
%U https://doi.org/10.18653/2026.findings-emnlp.2050
%P 43634-43645

Markdown (Informal)

[Notes to Self: Can LLMs Benefit from Experiential Abstractions?](https://aclanthology.org/2026.findings-emnlp.2050/) (Liu et al., Findings 2026)

• Notes to Self: Can LLMs Benefit from Experiential Abstractions? (Liu et al., Findings 2026)

ACL

• Chang Liu, Xinyu Li, and Artur Dubrawski. 2026. Notes to Self: Can LLMs Benefit from Experiential Abstractions?. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 43634–43645, Budapest, Hungary. Association for Computational Linguistics.