From Judgments to Issues: Structured Extraction of Legal Reasoning with Citation-Hallucination Control
Quick Answer
This paper shows that An automated pipeline utilizing DeepSeek V3 extracts structured legal issues from 330,000 Italian tax-court judgments, employing a hallucination-detection filter to ensure citation accuracy.
Quick Take
An automated pipeline utilizing DeepSeek V3 extracts structured legal issues from 330,000 Italian tax-court judgments, employing a hallucination-detection filter to ensure citation accuracy. This expert-validated system enhances legal reasoning applications, including issue-level retrieval and citation analysis.
Key Points
- Processes 330,000 Italian tax-court decisions with a cost-efficient model.
- Employs a hallucination-detection filter to validate legal citations.
- First expert-validated structured extraction pipeline for Italian tax law.
- Facilitates downstream applications like citation-network analysis.
- Validated on 50 judgments with inter-annotator agreement metrics.
Paper Resources
📖 Reader Mode
~2 min readAbstract:We present an automated pipeline that decomposes Italian tax-court judgments into individual legal issues and extracts, for each issue, a structured XML representation grounded in the IRAC framework and the legal syllogism. The pipeline targets a corpus of approximately $330{,}000$ first- and second-instance decisions of the Italian tax courts and is built around a capable yet cost-efficient general-purpose model (DeepSeek V3), a choice driven by the need to process several hundred thousand documents at a sustainable cost. To address the well-documented unreliability of large language models on legal citations, we couple the extraction step with an automatic hallucination-detection filter that compares the references produced by the model with those identified in the judgment text by a dedicated parser (Linkoln), normalised to standard identifiers (URN-NIR, ECLI, CELEX). We validate the pipeline on $50$ judgments annotated by two PhDs in tax law, computing inter-annotator agreement and LLM-vs-expert agreement on both issue extraction and legal citations, together with a stand-alone evaluation of the hallucination filter. To the best of our knowledge, this is the first issue-level, expert-validated structured extraction pipeline with hallucination control for Italian tax-court decisions, and it provides a concrete starting point for downstream applications such as issue-level retrieval, citation-network analysis, and the construction of large-scale datasets of legal reasoning.
| Comments: | 33 pages, 2 figures |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) |
| Cite as: | arXiv:2607.03325 [cs.CL] |
| (or arXiv:2607.03325v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2607.03325 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Giovanni Piccioli [view email]
[v1]
Fri, 3 Jul 2026 13:41:08 UTC (86 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →Quantifying Prior Dominance in Systems
The study introduces the Normalized Context Utilization (NCU) metric to evaluate Retrieval-Augmented Generation (RAG) systems, revealing that Small Language Models (SLMs) outperform larger models in factual extraction. The findings indicate that traditional scaling laws yield diminishing returns, with a commercial API frequently failing against adversarial evidence due to systemic confidence collapse.