@inproceedings{schumacher-etal-2026-prompting,
title = "Prompting Underestimates {LLM} Capability for Time Series Classification",
author = "Schumacher, Dan and
Nourbakhsh, Erfan and
Slavin, Rocky and
Rios, Anthony",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the {A}sia-Pacific Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.aacl-long.101/",
pages = "2212--2240",
ISBN = "979-8-89176-439-2",
abstract = "Prompt-based prediction pipelines suggest that large language models (LLMs) perform poorly on time-series classification, raising doubts about whether they encode meaningful temporal structure. We show that this conclusion reflects limitations of prompt-based generation rather than the model{'}s representational capacity by directly comparing prompt outputs with linear probes over the same internal representations while controlling for important confounding factors. While prompting performs near chance, linear probes improve average F1 from 0.15{--}0.26 to 0.61{--}0.67, often matching or exceeding specialized state-of-the-art (SoTA) time-series models. Layer-wise analyses further show that class-discriminative time-series information emerges in early transformer layers and is amplified with multimodal inputs and the presence of pretrained weights. Together, these results demonstrate a systematic mismatch between what LLMs internally represent and what prompt-based prediction pipelines reveal, leading current prompt-based prediction pipelines to underestimate the degree to which these models function as useful time-series feature extractors."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="schumacher-etal-2026-prompting">
<titleInfo>
<title>Prompting Underestimates LLM Capability for Time Series Classification</title>
</titleInfo>
<name type="personal">
<namePart type="given">Dan</namePart>
<namePart type="family">Schumacher</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Erfan</namePart>
<namePart type="family">Nourbakhsh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rocky</namePart>
<namePart type="family">Slavin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anthony</namePart>
<namePart type="family">Rios</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-439-2</identifier>
</relatedItem>
<abstract>Prompt-based prediction pipelines suggest that large language models (LLMs) perform poorly on time-series classification, raising doubts about whether they encode meaningful temporal structure. We show that this conclusion reflects limitations of prompt-based generation rather than the model’s representational capacity by directly comparing prompt outputs with linear probes over the same internal representations while controlling for important confounding factors. While prompting performs near chance, linear probes improve average F1 from 0.15–0.26 to 0.61–0.67, often matching or exceeding specialized state-of-the-art (SoTA) time-series models. Layer-wise analyses further show that class-discriminative time-series information emerges in early transformer layers and is amplified with multimodal inputs and the presence of pretrained weights. Together, these results demonstrate a systematic mismatch between what LLMs internally represent and what prompt-based prediction pipelines reveal, leading current prompt-based prediction pipelines to underestimate the degree to which these models function as useful time-series feature extractors.</abstract>
<identifier type="citekey">schumacher-etal-2026-prompting</identifier>
<location>
<url>https://aclanthology.org/2026.aacl-long.101/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>2212</start>
<end>2240</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Prompting Underestimates LLM Capability for Time Series Classification
%A Schumacher, Dan
%A Nourbakhsh, Erfan
%A Slavin, Rocky
%A Rios, Anthony
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-439-2
%F schumacher-etal-2026-prompting
%X Prompt-based prediction pipelines suggest that large language models (LLMs) perform poorly on time-series classification, raising doubts about whether they encode meaningful temporal structure. We show that this conclusion reflects limitations of prompt-based generation rather than the model’s representational capacity by directly comparing prompt outputs with linear probes over the same internal representations while controlling for important confounding factors. While prompting performs near chance, linear probes improve average F1 from 0.15–0.26 to 0.61–0.67, often matching or exceeding specialized state-of-the-art (SoTA) time-series models. Layer-wise analyses further show that class-discriminative time-series information emerges in early transformer layers and is amplified with multimodal inputs and the presence of pretrained weights. Together, these results demonstrate a systematic mismatch between what LLMs internally represent and what prompt-based prediction pipelines reveal, leading current prompt-based prediction pipelines to underestimate the degree to which these models function as useful time-series feature extractors.
%U https://aclanthology.org/2026.aacl-long.101/
%P 2212-2240
Markdown (Informal)
[Prompting Underestimates LLM Capability for Time Series Classification](https://aclanthology.org/2026.aacl-long.101/) (Schumacher et al., AACL-IJCNLP 2026)
ACL
- Dan Schumacher, Erfan Nourbakhsh, Rocky Slavin, and Anthony Rios. 2026. Prompting Underestimates LLM Capability for Time Series Classification. In Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2212–2240, Hengqin, China. Association for Computational Linguistics.