University of Twente Student Theses
As of Friday, 8 August 2025, the current Student Theses repository is no longer available for thesis uploads. A new Student Theses repository will be available starting Friday, 15 August 2025.
Legal Memorandum Generation Using Retrieval-Augmented Large Language Models and Dutch Case Law
Timoficiuc, Mihai (2025) Legal Memorandum Generation Using Retrieval-Augmented Large Language Models and Dutch Case Law.
|
PDF
1MB |
| Abstract: | Legal memo drafting is a critical but labor-intensive task in Dutch legal practice, requiring the synthesis of case law into structured, actionable documents. This thesis presents a Retrieval-Augmented Generation (RAG) system that generates legally grounded memoranda based on Dutch administrative court rulings and combines a curated corpus of 200 social security judgments with semantic chunking, pgvector-based retrieval, and GPT-4.1-driven memo generation. To enforce legal traceability, a structured intake form, and strict citation requirements were employed, combined with a modular evaluation framework developed to assess factual accuracy, citation precision, and semantic grounding using both automated heuristics and GPT-4.1 judgment. Results show the system achieves perfect citation precision (1.0), consistent recall (0.78), and F1 score (0.87) across multiple similarity thresholds. Reviewer LLMs added limited improvement, reinforcing the conclusion that robust retrieval and prompt design are more impactful than complex verification layers. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science, 86 law, 88 social and public administration |
| Programme: | Business & IT BSc (56066) |
| Link to this item: | https://purl.utwente.nl/essays/107362 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page