@inproceedings{gabryszak-etal-2026-facts,
title = "From Facts to Hypotheses: Joint Detection of Biomedical Relations and Epistemic Commitment Using {LLM}s",
author = {Gabryszak, Aleksandra and
Truong, Phuc Tran and
Binder, Arne and
Milosevic, Nikola and
Keese, Felix-Sebastian and
Rheinl{\"a}nder, Astrid and
Thomas, Philippe},
editor = "Piperidis, Stelios and
Bel, N{\'u}ria and
van den Heuvel, Henk and
Ide, Nancy and
Krek, Simon and
Toral, Antonio",
booktitle = "Proceedings of the Fifteenth Language Resources and Evaluation Conference",
month = may,
year = "2026",
address = "Palma de Mallorca, Spain",
publisher = "ELRA Language Resource Association",
url = "https://aclanthology.org/2026.lrec-1.602/",
doi = "10.63317/5prxfaguzng8",
pages = "7592--7605",
abstract = "Determining the factual status of biomedical statements, whether affirmed, negated, or uncertain, is essential for accurate understanding. To support research in this area, we introduce BioRelFact, a publicly available, expert-annotated dataset of 1,767 English biomedical sentences labeled with nine relation types and five levels of epistemic commitment. Using this dataset, we evaluate eight large language models (LLMs) from the GPT, Qwen, and Gemma families for joint relation extraction and epistemic classification. Among the evaluated models, GPT-OSS-20B performs best in both tasks (F1 77.3 for relation, 65.3 for commitment), followed by GPT-4o (75.9 and 60.2), while Qwen3-8B (Thinking) shows strong performance despite its smaller size (74.6 and 57.2). Domain adaptation has mixed effects: relative to their general-purpose counterparts, MedGemma-27B improves (+3.6 F1 for relation, +4.4 for factuality), whereas Qwen2.5-Aloe-Beta-7B declines ({--}4.3 and {--}3.5, respectively). Moreover, definition-based few-shot prompts consistently yield the best results for most models, and an explorative analysis of prediction errors suggests which specific linguistic features may drive model confusions."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="gabryszak-etal-2026-facts">
<titleInfo>
<title>From Facts to Hypotheses: Joint Detection of Biomedical Relations and Epistemic Commitment Using LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Aleksandra</namePart>
<namePart type="family">Gabryszak</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Phuc</namePart>
<namePart type="given">Tran</namePart>
<namePart type="family">Truong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arne</namePart>
<namePart type="family">Binder</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nikola</namePart>
<namePart type="family">Milosevic</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Felix-Sebastian</namePart>
<namePart type="family">Keese</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Astrid</namePart>
<namePart type="family">Rheinländer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Philippe</namePart>
<namePart type="family">Thomas</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 Fifteenth Language Resources and Evaluation Conference</title>
</titleInfo>
<name type="personal">
<namePart type="given">Stelios</namePart>
<namePart type="family">Piperidis</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Núria</namePart>
<namePart type="family">Bel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Henk</namePart>
<namePart type="family">van den Heuvel</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nancy</namePart>
<namePart type="family">Ide</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Simon</namePart>
<namePart type="family">Krek</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Antonio</namePart>
<namePart type="family">Toral</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resource Association</publisher>
<place>
<placeTerm type="text">Palma de Mallorca, Spain</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Determining the factual status of biomedical statements, whether affirmed, negated, or uncertain, is essential for accurate understanding. To support research in this area, we introduce BioRelFact, a publicly available, expert-annotated dataset of 1,767 English biomedical sentences labeled with nine relation types and five levels of epistemic commitment. Using this dataset, we evaluate eight large language models (LLMs) from the GPT, Qwen, and Gemma families for joint relation extraction and epistemic classification. Among the evaluated models, GPT-OSS-20B performs best in both tasks (F1 77.3 for relation, 65.3 for commitment), followed by GPT-4o (75.9 and 60.2), while Qwen3-8B (Thinking) shows strong performance despite its smaller size (74.6 and 57.2). Domain adaptation has mixed effects: relative to their general-purpose counterparts, MedGemma-27B improves (+3.6 F1 for relation, +4.4 for factuality), whereas Qwen2.5-Aloe-Beta-7B declines (–4.3 and –3.5, respectively). Moreover, definition-based few-shot prompts consistently yield the best results for most models, and an explorative analysis of prediction errors suggests which specific linguistic features may drive model confusions.</abstract>
<identifier type="citekey">gabryszak-etal-2026-facts</identifier>
<identifier type="doi">10.63317/5prxfaguzng8</identifier>
<location>
<url>https://aclanthology.org/2026.lrec-1.602/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>7592</start>
<end>7605</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T From Facts to Hypotheses: Joint Detection of Biomedical Relations and Epistemic Commitment Using LLMs
%A Gabryszak, Aleksandra
%A Truong, Phuc Tran
%A Binder, Arne
%A Milosevic, Nikola
%A Keese, Felix-Sebastian
%A Rheinländer, Astrid
%A Thomas, Philippe
%Y Piperidis, Stelios
%Y Bel, Núria
%Y van den Heuvel, Henk
%Y Ide, Nancy
%Y Krek, Simon
%Y Toral, Antonio
%S Proceedings of the Fifteenth Language Resources and Evaluation Conference
%D 2026
%8 May
%I ELRA Language Resource Association
%C Palma de Mallorca, Spain
%F gabryszak-etal-2026-facts
%X Determining the factual status of biomedical statements, whether affirmed, negated, or uncertain, is essential for accurate understanding. To support research in this area, we introduce BioRelFact, a publicly available, expert-annotated dataset of 1,767 English biomedical sentences labeled with nine relation types and five levels of epistemic commitment. Using this dataset, we evaluate eight large language models (LLMs) from the GPT, Qwen, and Gemma families for joint relation extraction and epistemic classification. Among the evaluated models, GPT-OSS-20B performs best in both tasks (F1 77.3 for relation, 65.3 for commitment), followed by GPT-4o (75.9 and 60.2), while Qwen3-8B (Thinking) shows strong performance despite its smaller size (74.6 and 57.2). Domain adaptation has mixed effects: relative to their general-purpose counterparts, MedGemma-27B improves (+3.6 F1 for relation, +4.4 for factuality), whereas Qwen2.5-Aloe-Beta-7B declines (–4.3 and –3.5, respectively). Moreover, definition-based few-shot prompts consistently yield the best results for most models, and an explorative analysis of prediction errors suggests which specific linguistic features may drive model confusions.
%R 10.63317/5prxfaguzng8
%U https://aclanthology.org/2026.lrec-1.602/
%U https://doi.org/10.63317/5prxfaguzng8
%P 7592-7605
Markdown (Informal)
[From Facts to Hypotheses: Joint Detection of Biomedical Relations and Epistemic Commitment Using LLMs](https://aclanthology.org/2026.lrec-1.602/) (Gabryszak et al., LREC 2026)
ACL