@inproceedings{xiao-etal-2025-longweave,
title = "{L}ong{W}eave: A Long-Form Generation Benchmark Bridging Real-World Relevance and Verifiability",
author = "Xiao, Zikai and
Huang, Fei and
Tu, Jianhong and
Wei, Jianhui and
Ma, Wen and
Zhou, Yuxuan and
Wu, Jian and
Yu, Bowen and
Liu, Zuozhu and
Lin, Junyang",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-emnlp.549/",
doi = "10.18653/v1/2025.findings-emnlp.549",
pages = "10386--10417",
ISBN = "979-8-89176-335-7",
abstract = "Generating long, informative, and factual outputs remains a major challenge for Large Language Models (LLMs). Existing benchmarks for long-form generation typically assess real-world queries with hard-to-verify metrics or use synthetic setups that ease evaluation but overlook real-world intricacies. In this paper, we introduce \textbf{LongWeave}, which balance real-world and verifiable assessment with Target-Anchored Evaluation (TAE). TAE constructs tasks by first defining verifiable targets within real-world scenarios, then systematically generating corresponding queries, textual materials, and anchors based on these targets. This ensures that tasks are both realistic and objectively assessable, enabling rigorous assessment of model capabilities in meeting complex real-world constraints. LongWeave supports customizable input/output lengths (up to 64K/8K tokens) across seven distinct tasks. Evaluation on 23 LLMs show that even state-of-the-art models encounter significant challenges in long-form generation as real-world complexity and output length increase. Dataset will be publicly available."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="xiao-etal-2025-longweave">
<titleInfo>
<title>LongWeave: A Long-Form Generation Benchmark Bridging Real-World Relevance and Verifiability</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zikai</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fei</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jianhong</namePart>
<namePart type="family">Tu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jianhui</namePart>
<namePart type="family">Wei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wen</namePart>
<namePart type="family">Ma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuxuan</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jian</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bowen</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zuozhu</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junyang</namePart>
<namePart type="family">Lin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Christos</namePart>
<namePart type="family">Christodoulopoulos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanmoy</namePart>
<namePart type="family">Chakraborty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Carolyn</namePart>
<namePart type="family">Rose</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Violet</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Suzhou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-335-7</identifier>
</relatedItem>
<abstract>Generating long, informative, and factual outputs remains a major challenge for Large Language Models (LLMs). Existing benchmarks for long-form generation typically assess real-world queries with hard-to-verify metrics or use synthetic setups that ease evaluation but overlook real-world intricacies. In this paper, we introduce LongWeave, which balance real-world and verifiable assessment with Target-Anchored Evaluation (TAE). TAE constructs tasks by first defining verifiable targets within real-world scenarios, then systematically generating corresponding queries, textual materials, and anchors based on these targets. This ensures that tasks are both realistic and objectively assessable, enabling rigorous assessment of model capabilities in meeting complex real-world constraints. LongWeave supports customizable input/output lengths (up to 64K/8K tokens) across seven distinct tasks. Evaluation on 23 LLMs show that even state-of-the-art models encounter significant challenges in long-form generation as real-world complexity and output length increase. Dataset will be publicly available.</abstract>
<identifier type="citekey">xiao-etal-2025-longweave</identifier>
<identifier type="doi">10.18653/v1/2025.findings-emnlp.549</identifier>
<location>
<url>https://aclanthology.org/2025.findings-emnlp.549/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>10386</start>
<end>10417</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T LongWeave: A Long-Form Generation Benchmark Bridging Real-World Relevance and Verifiability
%A Xiao, Zikai
%A Huang, Fei
%A Tu, Jianhong
%A Wei, Jianhui
%A Ma, Wen
%A Zhou, Yuxuan
%A Wu, Jian
%A Yu, Bowen
%A Liu, Zuozhu
%A Lin, Junyang
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Findings of the Association for Computational Linguistics: EMNLP 2025
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-335-7
%F xiao-etal-2025-longweave
%X Generating long, informative, and factual outputs remains a major challenge for Large Language Models (LLMs). Existing benchmarks for long-form generation typically assess real-world queries with hard-to-verify metrics or use synthetic setups that ease evaluation but overlook real-world intricacies. In this paper, we introduce LongWeave, which balance real-world and verifiable assessment with Target-Anchored Evaluation (TAE). TAE constructs tasks by first defining verifiable targets within real-world scenarios, then systematically generating corresponding queries, textual materials, and anchors based on these targets. This ensures that tasks are both realistic and objectively assessable, enabling rigorous assessment of model capabilities in meeting complex real-world constraints. LongWeave supports customizable input/output lengths (up to 64K/8K tokens) across seven distinct tasks. Evaluation on 23 LLMs show that even state-of-the-art models encounter significant challenges in long-form generation as real-world complexity and output length increase. Dataset will be publicly available.
%R 10.18653/v1/2025.findings-emnlp.549
%U https://aclanthology.org/2025.findings-emnlp.549/
%U https://doi.org/10.18653/v1/2025.findings-emnlp.549
%P 10386-10417
Markdown (Informal)
[LongWeave: A Long-Form Generation Benchmark Bridging Real-World Relevance and Verifiability](https://aclanthology.org/2025.findings-emnlp.549/) (Xiao et al., Findings 2025)
ACL
- Zikai Xiao, Fei Huang, Jianhong Tu, Jianhui Wei, Wen Ma, Yuxuan Zhou, Jian Wu, Bowen Yu, Zuozhu Liu, and Junyang Lin. 2025. LongWeave: A Long-Form Generation Benchmark Bridging Real-World Relevance and Verifiability. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 10386–10417, Suzhou, China. Association for Computational Linguistics.