@inproceedings{zhang-etal-2026-triggerbench,
title = "{T}rigger{B}ench: Investigating Prospective Memory for Large Language Models",
author = "Zhang, Tianhua and
Wang, Xinjiang and
Zhang, Qianxi and
Chen, Qi and
Li, Kun and
Chen, Yaoqi and
Wang, Dingdong and
Meng, Helen M. and
Lu, Yan",
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.1319/",
pages = "30034--30068",
ISBN = "979-8-89176-459-0",
abstract = "While Large Language Models (LLMs) are increasingly deployed in long interactions, existing evaluations focus predominantly on retrospective memory (RM) via explicit queries. Prospective memory (PM), the ability to spontaneously recall and act on latent constraints without direct requests, remains largely unevaluated. We introduce **TriggerBench**, a five-dimensional PM benchmark across daily assistants and professional workflows. TriggerBench pairs scenarios with matched RM controls, contrastive positive/negative variants, and overloaded triggers, enabling fine-grained measurement of proactive recall, false-alarm rate, and attentional robustness. We find: (i) PM shows a precision-recall trade-off and attentional fragility. Though enhanced reasoning improves proactive recall, models can overfit to always-remind heuristics. Furthermore, PM accuracy degrades substantially under implicit constraints or triggers overloaded by concurrent requests, indicating that robust PM remains an open challenge. (ii) PM is notably harder than RM: on identical contexts, RM remains near-saturated up to 100K tokens, while PM decays sharply as context length increases. (iii) PM may serve as a behavioral probe of spare reasoning capacity. Pairing PM with AIME-2025 math problems reveals that successful trajectories yield higher PM accuracy than failed ones at the same context length, showing PM is a behavioral correlate of spare reasoning budget that token count obscures. Project page: https://github.com/KristenZHANG/TriggerBench-Official."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2026-triggerbench">
<titleInfo>
<title>TriggerBench: Investigating Prospective Memory for Large Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Tianhua</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinjiang</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qianxi</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qi</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kun</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yaoqi</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dingdong</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Helen</namePart>
<namePart type="given">M</namePart>
<namePart type="family">Meng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yan</namePart>
<namePart type="family">Lu</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>While Large Language Models (LLMs) are increasingly deployed in long interactions, existing evaluations focus predominantly on retrospective memory (RM) via explicit queries. Prospective memory (PM), the ability to spontaneously recall and act on latent constraints without direct requests, remains largely unevaluated. We introduce **TriggerBench**, a five-dimensional PM benchmark across daily assistants and professional workflows. TriggerBench pairs scenarios with matched RM controls, contrastive positive/negative variants, and overloaded triggers, enabling fine-grained measurement of proactive recall, false-alarm rate, and attentional robustness. We find: (i) PM shows a precision-recall trade-off and attentional fragility. Though enhanced reasoning improves proactive recall, models can overfit to always-remind heuristics. Furthermore, PM accuracy degrades substantially under implicit constraints or triggers overloaded by concurrent requests, indicating that robust PM remains an open challenge. (ii) PM is notably harder than RM: on identical contexts, RM remains near-saturated up to 100K tokens, while PM decays sharply as context length increases. (iii) PM may serve as a behavioral probe of spare reasoning capacity. Pairing PM with AIME-2025 math problems reveals that successful trajectories yield higher PM accuracy than failed ones at the same context length, showing PM is a behavioral correlate of spare reasoning budget that token count obscures. Project page: https://github.com/KristenZHANG/TriggerBench-Official.</abstract>
<identifier type="citekey">zhang-etal-2026-triggerbench</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1319/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>30034</start>
<end>30068</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T TriggerBench: Investigating Prospective Memory for Large Language Models
%A Zhang, Tianhua
%A Wang, Xinjiang
%A Zhang, Qianxi
%A Chen, Qi
%A Li, Kun
%A Chen, Yaoqi
%A Wang, Dingdong
%A Meng, Helen M.
%A Lu, Yan
%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 zhang-etal-2026-triggerbench
%X While Large Language Models (LLMs) are increasingly deployed in long interactions, existing evaluations focus predominantly on retrospective memory (RM) via explicit queries. Prospective memory (PM), the ability to spontaneously recall and act on latent constraints without direct requests, remains largely unevaluated. We introduce **TriggerBench**, a five-dimensional PM benchmark across daily assistants and professional workflows. TriggerBench pairs scenarios with matched RM controls, contrastive positive/negative variants, and overloaded triggers, enabling fine-grained measurement of proactive recall, false-alarm rate, and attentional robustness. We find: (i) PM shows a precision-recall trade-off and attentional fragility. Though enhanced reasoning improves proactive recall, models can overfit to always-remind heuristics. Furthermore, PM accuracy degrades substantially under implicit constraints or triggers overloaded by concurrent requests, indicating that robust PM remains an open challenge. (ii) PM is notably harder than RM: on identical contexts, RM remains near-saturated up to 100K tokens, while PM decays sharply as context length increases. (iii) PM may serve as a behavioral probe of spare reasoning capacity. Pairing PM with AIME-2025 math problems reveals that successful trajectories yield higher PM accuracy than failed ones at the same context length, showing PM is a behavioral correlate of spare reasoning budget that token count obscures. Project page: https://github.com/KristenZHANG/TriggerBench-Official.
%U https://aclanthology.org/2026.emnlp-main.1319/
%P 30034-30068
Markdown (Informal)
[TriggerBench: Investigating Prospective Memory for Large Language Models](https://aclanthology.org/2026.emnlp-main.1319/) (Zhang et al., EMNLP 2026)
ACL
- Tianhua Zhang, Xinjiang Wang, Qianxi Zhang, Qi Chen, Kun Li, Yaoqi Chen, Dingdong Wang, Helen M. Meng, and Yan Lu. 2026. TriggerBench: Investigating Prospective Memory for Large Language Models. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 30034–30068, Budapest, Hungary. Association for Computational Linguistics.