@inproceedings{yan-etal-2026-specification,
title = "When the Specification Emerges: Benchmarking Faithfulness Loss in Long-Horizon Coding Agents",
author = "Yan, Lu and
Chen, Xuan and
Zhang, Xiangyu",
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.76/",
pages = "1552--1571",
ISBN = "979-8-89176-462-0",
abstract = "Current coding-agent benchmarks usually provide the full task specification upfront. Research-code implementation may instead re- quire an agent to preserve design commitments disclosed progressively across a long session. We introduce SLUMP, a controlled benchmark for testing implementation faithfulness when a mathematically structured machine-learning target is disclosed through synthetic, paper- grounded multi-turn interactions rather than in a single canonical implementation specification. The benchmark contains 20 recent ML papers, 371 atomic verifiable components, and approx- imately 60 coding requests per paper. An expo- sure audit verifies whether scored components are recoverable from the visible interaction. On Claude Code and Codex, the single-shot control achieves higher overall implementation fidelity on 16/20 and 14/20 papers, respectively. Struc- tural integration degrades under progressive dis- closure on both platforms, while semantic com- ponent faithfulness degrades substantially on Claude Code and remains approximately sta- ble on Codex. These conclusions persist under stricter exposure filters and a targeted human correction of judge false-absence errors. As a mitigation case study, ProjectGuard main- tains external semantic and structural project state and recovers 90{\%} of Claude Code{'}s mea- sured faithfulness gap. Code and benchmark artifacts are available at https://github.com/lunaryan/drift-benchmark."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yan-etal-2026-specification">
<titleInfo>
<title>When the Specification Emerges: Benchmarking Faithfulness Loss in Long-Horizon Coding Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lu</namePart>
<namePart type="family">Yan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xuan</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiangyu</namePart>
<namePart type="family">Zhang</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>Current coding-agent benchmarks usually provide the full task specification upfront. Research-code implementation may instead re- quire an agent to preserve design commitments disclosed progressively across a long session. We introduce SLUMP, a controlled benchmark for testing implementation faithfulness when a mathematically structured machine-learning target is disclosed through synthetic, paper- grounded multi-turn interactions rather than in a single canonical implementation specification. The benchmark contains 20 recent ML papers, 371 atomic verifiable components, and approx- imately 60 coding requests per paper. An expo- sure audit verifies whether scored components are recoverable from the visible interaction. On Claude Code and Codex, the single-shot control achieves higher overall implementation fidelity on 16/20 and 14/20 papers, respectively. Struc- tural integration degrades under progressive dis- closure on both platforms, while semantic com- ponent faithfulness degrades substantially on Claude Code and remains approximately sta- ble on Codex. These conclusions persist under stricter exposure filters and a targeted human correction of judge false-absence errors. As a mitigation case study, ProjectGuard main- tains external semantic and structural project state and recovers 90% of Claude Code’s mea- sured faithfulness gap. Code and benchmark artifacts are available at https://github.com/lunaryan/drift-benchmark.</abstract>
<identifier type="citekey">yan-etal-2026-specification</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.76/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1552</start>
<end>1571</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T When the Specification Emerges: Benchmarking Faithfulness Loss in Long-Horizon Coding Agents
%A Yan, Lu
%A Chen, Xuan
%A Zhang, Xiangyu
%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 yan-etal-2026-specification
%X Current coding-agent benchmarks usually provide the full task specification upfront. Research-code implementation may instead re- quire an agent to preserve design commitments disclosed progressively across a long session. We introduce SLUMP, a controlled benchmark for testing implementation faithfulness when a mathematically structured machine-learning target is disclosed through synthetic, paper- grounded multi-turn interactions rather than in a single canonical implementation specification. The benchmark contains 20 recent ML papers, 371 atomic verifiable components, and approx- imately 60 coding requests per paper. An expo- sure audit verifies whether scored components are recoverable from the visible interaction. On Claude Code and Codex, the single-shot control achieves higher overall implementation fidelity on 16/20 and 14/20 papers, respectively. Struc- tural integration degrades under progressive dis- closure on both platforms, while semantic com- ponent faithfulness degrades substantially on Claude Code and remains approximately sta- ble on Codex. These conclusions persist under stricter exposure filters and a targeted human correction of judge false-absence errors. As a mitigation case study, ProjectGuard main- tains external semantic and structural project state and recovers 90% of Claude Code’s mea- sured faithfulness gap. Code and benchmark artifacts are available at https://github.com/lunaryan/drift-benchmark.
%U https://aclanthology.org/2026.findings-emnlp.76/
%P 1552-1571
Markdown (Informal)
[When the Specification Emerges: Benchmarking Faithfulness Loss in Long-Horizon Coding Agents](https://aclanthology.org/2026.findings-emnlp.76/) (Yan et al., Findings 2026)
ACL