@inproceedings{schomacker-etal-2026-revisiting,
title = "Revisiting {G}erman Complex Word Identification: Contextualized {LLM}s and Feature Injection",
author = "Schomacker, Thorben and
Yimam, Seid Muhie and
Biemann, Chris and
Tropmann-Frick, Marina",
editor = "Shardlow, Matthew and
Fran{\c{c}}ois, Thomas and
Amaro, Raquel and
Baptista, Jorge and
Cardon, R{\'e}mi and
Ribeiro, Eug{\'e}nio and
Saggion, Horacio and
Stodden, Regina and
Todirascu, Amalia and
Wilkens, Rodrigo",
booktitle = "Proceedings of the Joint Workshop on Readability and Text Simplification ({READI}x{TSAR}) @ {LREC} 2026",
month = may,
year = "2026",
address = "Palma, Mallorca (Spain)",
publisher = "ELRA Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.readi-1.1/",
doi = "10.63317/4z33nthezmjs",
pages = "1--11",
abstract = "Complex word identification (CWI) is essential in text simplification, yet work on German CWI remains comparatively limited. To address this gap, we investigate the capabilities of three state-of-the-art LLMs and compare them to previously proposed baseline systems. We fine-tune the LLMs in three setups: (i) using the target expression only, (ii) using the target expression together with its sentence-level context, and (iii) using the context and injection of classical machine learning features. Our results show that while pretrained-only LLMs fall short, fine-tuned LLMs set new benchmarks for both binary and probabilistic CWI. In addition, embedding the target in its context sentence improves performance, whereas feature injection has no clearly measurable effect. All models in this paper are trained on the probabilistic CWI task and additionally evaluated on the binary task; thus, we publish a single model that supports both evaluation views We released all accompanying resources (\url{https://github.com/tschomacker/german-cwi-llm}) and model checkpoints (\url{https://huggingface.co/collections/tschomacker/german-cwi-llm})."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="schomacker-etal-2026-revisiting">
<titleInfo>
<title>Revisiting German Complex Word Identification: Contextualized LLMs and Feature Injection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Thorben</namePart>
<namePart type="family">Schomacker</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Seid</namePart>
<namePart type="given">Muhie</namePart>
<namePart type="family">Yimam</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chris</namePart>
<namePart type="family">Biemann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marina</namePart>
<namePart type="family">Tropmann-Frick</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Joint Workshop on Readability and Text Simplification (READIxTSAR) @ LREC 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Matthew</namePart>
<namePart type="family">Shardlow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Thomas</namePart>
<namePart type="family">François</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Raquel</namePart>
<namePart type="family">Amaro</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jorge</namePart>
<namePart type="family">Baptista</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rémi</namePart>
<namePart type="family">Cardon</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eugénio</namePart>
<namePart type="family">Ribeiro</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Horacio</namePart>
<namePart type="family">Saggion</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Regina</namePart>
<namePart type="family">Stodden</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Amalia</namePart>
<namePart type="family">Todirascu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rodrigo</namePart>
<namePart type="family">Wilkens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resources Association (ELRA)</publisher>
<place>
<placeTerm type="text">Palma, Mallorca (Spain)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Complex word identification (CWI) is essential in text simplification, yet work on German CWI remains comparatively limited. To address this gap, we investigate the capabilities of three state-of-the-art LLMs and compare them to previously proposed baseline systems. We fine-tune the LLMs in three setups: (i) using the target expression only, (ii) using the target expression together with its sentence-level context, and (iii) using the context and injection of classical machine learning features. Our results show that while pretrained-only LLMs fall short, fine-tuned LLMs set new benchmarks for both binary and probabilistic CWI. In addition, embedding the target in its context sentence improves performance, whereas feature injection has no clearly measurable effect. All models in this paper are trained on the probabilistic CWI task and additionally evaluated on the binary task; thus, we publish a single model that supports both evaluation views We released all accompanying resources (https://github.com/tschomacker/german-cwi-llm) and model checkpoints (https://huggingface.co/collections/tschomacker/german-cwi-llm).</abstract>
<identifier type="citekey">schomacker-etal-2026-revisiting</identifier>
<identifier type="doi">10.63317/4z33nthezmjs</identifier>
<location>
<url>https://aclanthology.org/2026.readi-1.1/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>1</start>
<end>11</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Revisiting German Complex Word Identification: Contextualized LLMs and Feature Injection
%A Schomacker, Thorben
%A Yimam, Seid Muhie
%A Biemann, Chris
%A Tropmann-Frick, Marina
%Y Shardlow, Matthew
%Y François, Thomas
%Y Amaro, Raquel
%Y Baptista, Jorge
%Y Cardon, Rémi
%Y Ribeiro, Eugénio
%Y Saggion, Horacio
%Y Stodden, Regina
%Y Todirascu, Amalia
%Y Wilkens, Rodrigo
%S Proceedings of the Joint Workshop on Readability and Text Simplification (READIxTSAR) @ LREC 2026
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma, Mallorca (Spain)
%F schomacker-etal-2026-revisiting
%X Complex word identification (CWI) is essential in text simplification, yet work on German CWI remains comparatively limited. To address this gap, we investigate the capabilities of three state-of-the-art LLMs and compare them to previously proposed baseline systems. We fine-tune the LLMs in three setups: (i) using the target expression only, (ii) using the target expression together with its sentence-level context, and (iii) using the context and injection of classical machine learning features. Our results show that while pretrained-only LLMs fall short, fine-tuned LLMs set new benchmarks for both binary and probabilistic CWI. In addition, embedding the target in its context sentence improves performance, whereas feature injection has no clearly measurable effect. All models in this paper are trained on the probabilistic CWI task and additionally evaluated on the binary task; thus, we publish a single model that supports both evaluation views We released all accompanying resources (https://github.com/tschomacker/german-cwi-llm) and model checkpoints (https://huggingface.co/collections/tschomacker/german-cwi-llm).
%R 10.63317/4z33nthezmjs
%U https://aclanthology.org/2026.readi-1.1/
%U https://doi.org/10.63317/4z33nthezmjs
%P 1-11
Markdown (Informal)
[Revisiting German Complex Word Identification: Contextualized LLMs and Feature Injection](https://aclanthology.org/2026.readi-1.1/) (Schomacker et al., READI-TSAR 2026)
ACL