@inproceedings{qi-etal-2026-compile,
title = "Compile, Don{'}t Memorize: A Context Compilation Architecture ({CCA}) for In-Context Learning",
author = "Qi, Jinhu and
Hu, Minda and
Zhang, Wentao and
Jin, Weiqiang and
Chen, Yanyu and
Wang, Junli and
King, Irwin",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1123/",
pages = "23781--23803",
ISBN = "979-8-89176-462-0",
abstract = "Large language models (LLMs) increasingly handle in-context learning (ICL) tasks where a long, novel context defines the rules, knowledge, and output schema for a series of questions. On benchmarks that grade against every detail of the context, even strong open-weights models pass only 12-16{\%} of tasks: a single overlooked rule fails the whole response. We argue this brittleness is structural{---}the dominant ``read-and-reason'' paradigm asks the model to extract, plan, generate, and self-verify in one forward pass{---}and ask whether explicit context compilation can fix it, how it compares to existing long-context strategies (gist retrieval, multi-agent self-play), and where the resulting harness benefit holds across task structure and model scale. We propose the Context Compilation Architecture (CCA): compile the given context into a structured IR of rules, schemas, and exact terms; generate executable verifiers from the IR; run a checked reasoning loop that corrects any violation those verifiers flag. On CL-bench (1,899 tasks across 4 open base models), CCA outperforms vanilla prompting and two long-context baselines (ReadAgent-P, Ctx2Skill) on every base model, lifting Kimi K2.5 from 15.4{\%} to 21.4{\%} with gains concentrated on rule-dense sub-categories."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="qi-etal-2026-compile">
<titleInfo>
<title>Compile, Don’t Memorize: A Context Compilation Architecture (CCA) for In-Context Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jinhu</namePart>
<namePart type="family">Qi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Minda</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wentao</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weiqiang</namePart>
<namePart type="family">Jin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yanyu</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junli</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Irwin</namePart>
<namePart type="family">King</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>Large language models (LLMs) increasingly handle in-context learning (ICL) tasks where a long, novel context defines the rules, knowledge, and output schema for a series of questions. On benchmarks that grade against every detail of the context, even strong open-weights models pass only 12-16% of tasks: a single overlooked rule fails the whole response. We argue this brittleness is structural—the dominant “read-and-reason” paradigm asks the model to extract, plan, generate, and self-verify in one forward pass—and ask whether explicit context compilation can fix it, how it compares to existing long-context strategies (gist retrieval, multi-agent self-play), and where the resulting harness benefit holds across task structure and model scale. We propose the Context Compilation Architecture (CCA): compile the given context into a structured IR of rules, schemas, and exact terms; generate executable verifiers from the IR; run a checked reasoning loop that corrects any violation those verifiers flag. On CL-bench (1,899 tasks across 4 open base models), CCA outperforms vanilla prompting and two long-context baselines (ReadAgent-P, Ctx2Skill) on every base model, lifting Kimi K2.5 from 15.4% to 21.4% with gains concentrated on rule-dense sub-categories.</abstract>
<identifier type="citekey">qi-etal-2026-compile</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1123/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>23781</start>
<end>23803</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Compile, Don’t Memorize: A Context Compilation Architecture (CCA) for In-Context Learning
%A Qi, Jinhu
%A Hu, Minda
%A Zhang, Wentao
%A Jin, Weiqiang
%A Chen, Yanyu
%A Wang, Junli
%A King, Irwin
%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 qi-etal-2026-compile
%X Large language models (LLMs) increasingly handle in-context learning (ICL) tasks where a long, novel context defines the rules, knowledge, and output schema for a series of questions. On benchmarks that grade against every detail of the context, even strong open-weights models pass only 12-16% of tasks: a single overlooked rule fails the whole response. We argue this brittleness is structural—the dominant “read-and-reason” paradigm asks the model to extract, plan, generate, and self-verify in one forward pass—and ask whether explicit context compilation can fix it, how it compares to existing long-context strategies (gist retrieval, multi-agent self-play), and where the resulting harness benefit holds across task structure and model scale. We propose the Context Compilation Architecture (CCA): compile the given context into a structured IR of rules, schemas, and exact terms; generate executable verifiers from the IR; run a checked reasoning loop that corrects any violation those verifiers flag. On CL-bench (1,899 tasks across 4 open base models), CCA outperforms vanilla prompting and two long-context baselines (ReadAgent-P, Ctx2Skill) on every base model, lifting Kimi K2.5 from 15.4% to 21.4% with gains concentrated on rule-dense sub-categories.
%U https://aclanthology.org/2026.findings-emnlp.1123/
%P 23781-23803
Markdown (Informal)
[Compile, Don’t Memorize: A Context Compilation Architecture (CCA) for In-Context Learning](https://aclanthology.org/2026.findings-emnlp.1123/) (Qi et al., Findings 2026)
ACL
- Jinhu Qi, Minda Hu, Wentao Zhang, Weiqiang Jin, Yanyu Chen, Junli Wang, and Irwin King. 2026. Compile, Don’t Memorize: A Context Compilation Architecture (CCA) for In-Context Learning. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 23781–23803, Budapest, Hungary. Association for Computational Linguistics.