@inproceedings{qin-etal-2026-llm,
title = "What Are {LLM} Agents Actually Learning on Low-Budget Molecular Optimization Benchmarks?",
author = "Qin, Jiayu and
Luo, Zhengquan and
Chen, Changyou and
xu, Zhiqiang",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2447/",
pages = "55505--55516",
ISBN = "979-8-89176-459-0",
abstract = "Recent work reports large gains for LLM agents on low-budget molecular optimization, with 2-3 times AUC improvements over reinforcement-learning and genetic-algorithm baselines on the PMO benchmark and similar gains on protein-targeted docking benchmarks. We investigate whether these gains reflect genuine optimization or benchmark artifacts, by re-evaluating a representative LLM agent under controlled blinding protocols and against a trivial random-walk baseline. We find that on PMO at the 50-call budget, the agent{'}s apparent advantage collapses below classical baselines once task names and curated reference compounds are removed from its prompt; a 2 times 2 ablation shows that retrieval from a small reference dictionary, not iterative optimization, is the dominant driver of the headline number. On a new protein-targeted lead-optimization benchmark with an AutoDock Vina composite oracle, a single-edit halogen/methyl random-walk mutator outperforms the LLM agent, a graph-based GA, and a non-iterative GPT baseline on every aggregate metric{---}a property of the oracle landscape under tight budgets, not of LLM agents per se. We release both protocols (name-blinded PMO scoring; random-walk lower bound for Vina-based protein benchmarks) as standard sanity checks for future LLM-agent papers in molecular optimization."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="qin-etal-2026-llm">
<titleInfo>
<title>What Are LLM Agents Actually Learning on Low-Budget Molecular Optimization Benchmarks?</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jiayu</namePart>
<namePart type="family">Qin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhengquan</namePart>
<namePart type="family">Luo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Changyou</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhiqiang</namePart>
<namePart type="family">xu</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>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</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">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Recent work reports large gains for LLM agents on low-budget molecular optimization, with 2-3 times AUC improvements over reinforcement-learning and genetic-algorithm baselines on the PMO benchmark and similar gains on protein-targeted docking benchmarks. We investigate whether these gains reflect genuine optimization or benchmark artifacts, by re-evaluating a representative LLM agent under controlled blinding protocols and against a trivial random-walk baseline. We find that on PMO at the 50-call budget, the agent’s apparent advantage collapses below classical baselines once task names and curated reference compounds are removed from its prompt; a 2 times 2 ablation shows that retrieval from a small reference dictionary, not iterative optimization, is the dominant driver of the headline number. On a new protein-targeted lead-optimization benchmark with an AutoDock Vina composite oracle, a single-edit halogen/methyl random-walk mutator outperforms the LLM agent, a graph-based GA, and a non-iterative GPT baseline on every aggregate metric—a property of the oracle landscape under tight budgets, not of LLM agents per se. We release both protocols (name-blinded PMO scoring; random-walk lower bound for Vina-based protein benchmarks) as standard sanity checks for future LLM-agent papers in molecular optimization.</abstract>
<identifier type="citekey">qin-etal-2026-llm</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2447/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>55505</start>
<end>55516</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T What Are LLM Agents Actually Learning on Low-Budget Molecular Optimization Benchmarks?
%A Qin, Jiayu
%A Luo, Zhengquan
%A Chen, Changyou
%A xu, Zhiqiang
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F qin-etal-2026-llm
%X Recent work reports large gains for LLM agents on low-budget molecular optimization, with 2-3 times AUC improvements over reinforcement-learning and genetic-algorithm baselines on the PMO benchmark and similar gains on protein-targeted docking benchmarks. We investigate whether these gains reflect genuine optimization or benchmark artifacts, by re-evaluating a representative LLM agent under controlled blinding protocols and against a trivial random-walk baseline. We find that on PMO at the 50-call budget, the agent’s apparent advantage collapses below classical baselines once task names and curated reference compounds are removed from its prompt; a 2 times 2 ablation shows that retrieval from a small reference dictionary, not iterative optimization, is the dominant driver of the headline number. On a new protein-targeted lead-optimization benchmark with an AutoDock Vina composite oracle, a single-edit halogen/methyl random-walk mutator outperforms the LLM agent, a graph-based GA, and a non-iterative GPT baseline on every aggregate metric—a property of the oracle landscape under tight budgets, not of LLM agents per se. We release both protocols (name-blinded PMO scoring; random-walk lower bound for Vina-based protein benchmarks) as standard sanity checks for future LLM-agent papers in molecular optimization.
%U https://aclanthology.org/2026.emnlp-main.2447/
%P 55505-55516
Markdown (Informal)
[What Are LLM Agents Actually Learning on Low-Budget Molecular Optimization Benchmarks?](https://aclanthology.org/2026.emnlp-main.2447/) (Qin et al., EMNLP 2026)
ACL