LLM-based syntactic parsing: impact of generating formats
Received: 09 Nov 2026 Revised: 10 Nov 2026 Accepted: 10 Nov 2026
Published: 2026, vol. 30, issue 4, pp. 445–448
Abstract
This article studies generative LLM-based syntactic parsing. The impact of bracketed syntactic representation formats (GRCT and LCT) on parsing quality is investigated. Research has established that an LLM can be successfully fine-tuned for analysis on each of the formats. For both formats, according to the LAS_LLM metric, the fine-tuned Qwen3-4B model exceeds the classic neural network parsers Stanza and DeepPavlov. Moreover, the fine-tuned LLM is resistant to domain shift.
Keywords: parsing, fine-tuning, large language models, dependency tree, bracketed format.
BibTeX
@article{IS-Shamaeva-Loukachevitch2026,
author = {Shamaeva, Elena Denisovna and Loukachevitch, Natalia Valentinovna},
title = {{LLM-based syntactic parsing: impact of generating formats}},
journal = {Intelligent Systems. Theory and Applications},
year = {2026},
volume = {30},
number = {4},
pages = {445--448},
}
AMSBIB
\Bibitem{IS-Shamaeva-Loukachevitch2026}
\by E.\,D.~Shamaeva, N.\,V.~Loukachevitch
\paper LLM-based syntactic parsing: impact of generating formats
\jour Intelligent Systems. Theory and Applications
\yr 2026
\vol 30
\issue 4
\pages 445--448
\lang In Russian
RU