@inproceedings{chang-chen-2026-users,
title = "When Users Don{'}t Ask: Benchmarking Context-Driven Memory Retrieval in Conversational Agents",
author = "Chang, Wen Yu and
Chen, Yun-Nung",
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.1289/",
doi = "10.18653/2026.findings-emnlp.1289",
pages = "27273--27293",
ISBN = "979-8-89176-462-0",
abstract = "Large Language Models (LLMs) are increasingly deployed as long-horizon conversational agents, motivating growing interest in memory systems. However, existing benchmarks primarily evaluate memory through QA-style probing rather than in-situ conversational usage. We introduce LOCOMO-CONV, a conversational memory benchmark derived from LoCoMo with four query styles: dialog, implicit, counterfactual, and composed. Across four representative memory systems, we evaluate both retrieval recall and end-to-end response quality. Our experiments show that conversational framing exposes substantial retrieval gaps overlooked by QA benchmarks, especially on implicit and composed queries, which multi-facet query rewriting narrows for raw-turn but not abstractive memory. We further find that strong retrieval does not fully translate into response quality, and that implicit queries exhibit silent grounding, where memory improves contextual grounding without explicitly surfacing the gold fact. These results point to reasoning-based memory elaboration as a promising direction, and we release auxiliary supportive{\_}memory annotations capturing conversationally useful context beyond the original gold evidence."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chang-chen-2026-users">
<titleInfo>
<title>When Users Don’t Ask: Benchmarking Context-Driven Memory Retrieval in Conversational Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wen</namePart>
<namePart type="given">Yu</namePart>
<namePart type="family">Chang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yun-Nung</namePart>
<namePart type="family">Chen</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) are increasingly deployed as long-horizon conversational agents, motivating growing interest in memory systems. However, existing benchmarks primarily evaluate memory through QA-style probing rather than in-situ conversational usage. We introduce LOCOMO-CONV, a conversational memory benchmark derived from LoCoMo with four query styles: dialog, implicit, counterfactual, and composed. Across four representative memory systems, we evaluate both retrieval recall and end-to-end response quality. Our experiments show that conversational framing exposes substantial retrieval gaps overlooked by QA benchmarks, especially on implicit and composed queries, which multi-facet query rewriting narrows for raw-turn but not abstractive memory. We further find that strong retrieval does not fully translate into response quality, and that implicit queries exhibit silent grounding, where memory improves contextual grounding without explicitly surfacing the gold fact. These results point to reasoning-based memory elaboration as a promising direction, and we release auxiliary supportive_memory annotations capturing conversationally useful context beyond the original gold evidence.</abstract>
<identifier type="citekey">chang-chen-2026-users</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1289</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1289/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>27273</start>
<end>27293</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T When Users Don’t Ask: Benchmarking Context-Driven Memory Retrieval in Conversational Agents
%A Chang, Wen Yu
%A Chen, Yun-Nung
%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 chang-chen-2026-users
%X Large Language Models (LLMs) are increasingly deployed as long-horizon conversational agents, motivating growing interest in memory systems. However, existing benchmarks primarily evaluate memory through QA-style probing rather than in-situ conversational usage. We introduce LOCOMO-CONV, a conversational memory benchmark derived from LoCoMo with four query styles: dialog, implicit, counterfactual, and composed. Across four representative memory systems, we evaluate both retrieval recall and end-to-end response quality. Our experiments show that conversational framing exposes substantial retrieval gaps overlooked by QA benchmarks, especially on implicit and composed queries, which multi-facet query rewriting narrows for raw-turn but not abstractive memory. We further find that strong retrieval does not fully translate into response quality, and that implicit queries exhibit silent grounding, where memory improves contextual grounding without explicitly surfacing the gold fact. These results point to reasoning-based memory elaboration as a promising direction, and we release auxiliary supportive_memory annotations capturing conversationally useful context beyond the original gold evidence.
%R 10.18653/2026.findings-emnlp.1289
%U https://aclanthology.org/2026.findings-emnlp.1289/
%U https://doi.org/10.18653/2026.findings-emnlp.1289
%P 27273-27293
Markdown (Informal)
[When Users Don’t Ask: Benchmarking Context-Driven Memory Retrieval in Conversational Agents](https://aclanthology.org/2026.findings-emnlp.1289/) (Chang & Chen, Findings 2026)
ACL