Research

Paper

AI LLM March 02, 2026

nchellwig at SemEval-2026 Task 3: Self-Consistent Structured Generation (SCSG) for Dimensional Aspect-Based Sentiment Analysis using Large Language Models

Authors

Nils Constantin Hellwig, Jakob Fehle, Udo Kruschwitz, Christian Wolff

Abstract

We present Self-Consistent Structured Generation (SCSG) for Dimensional Aspect-Based Sentiment Analysis in SemEval-2026 Task 3 (Track A). SCSG enhances prediction reliability by executing a LoRA-adapted large language model multiple times per instance, retaining only tuples that achieve a majority consensus across runs. To mitigate the computational overhead of multiple forward passes, we leverage vLLM's PagedAttention mechanism for efficient key--value cache reuse. Evaluation across 6 languages and 8 language--domain combinations demonstrates that self-consistency with 15 executions yields statistically significant improvements over single-inference prompting, with our system (leveraging Gemma 3) ranking in the top seven across all settings, achieving second place on three out of four English subsets and first place on Tatar-Restaurant for DimASTE.

Metadata

arXiv ID: 2603.01788
Provider: ARXIV
Primary Category: cs.CL
Published: 2026-03-02
Fetched: 2026-03-03 04:34

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.01788v1</id>\n    <title>nchellwig at SemEval-2026 Task 3: Self-Consistent Structured Generation (SCSG) for Dimensional Aspect-Based Sentiment Analysis using Large Language Models</title>\n    <updated>2026-03-02T12:13:03Z</updated>\n    <link href='https://arxiv.org/abs/2603.01788v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.01788v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>We present Self-Consistent Structured Generation (SCSG) for Dimensional Aspect-Based Sentiment Analysis in SemEval-2026 Task 3 (Track A). SCSG enhances prediction reliability by executing a LoRA-adapted large language model multiple times per instance, retaining only tuples that achieve a majority consensus across runs. To mitigate the computational overhead of multiple forward passes, we leverage vLLM's PagedAttention mechanism for efficient key--value cache reuse. Evaluation across 6 languages and 8 language--domain combinations demonstrates that self-consistency with 15 executions yields statistically significant improvements over single-inference prompting, with our system (leveraging Gemma 3) ranking in the top seven across all settings, achieving second place on three out of four English subsets and first place on Tatar-Restaurant for DimASTE.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <published>2026-03-02T12:13:03Z</published>\n    <arxiv:primary_category term='cs.CL'/>\n    <author>\n      <name>Nils Constantin Hellwig</name>\n    </author>\n    <author>\n      <name>Jakob Fehle</name>\n    </author>\n    <author>\n      <name>Udo Kruschwitz</name>\n    </author>\n    <author>\n      <name>Christian Wolff</name>\n    </author>\n  </entry>"
}