LegalCiteBench reveals that current LLMs achieve under 7% accuracy on closed-book legal citation retrieval and completion tasks, with misleading answer rates above 94% for nearly all tested models.
arXiv preprint arXiv:2509.09969 , year=
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 3years
2026 3verdicts
UNVERDICTED 3roles
background 1polarities
background 1representative citing papers
Lightweight LLMs are benchmarked for court view generation and charge prediction across architectures, sizes, DNN comparisons, and task ordering on three datasets using the new CVGEvalKit framework.
A 1D CNN with FastText embeddings classifies legal texts at 97.26% accuracy using 5.1 million parameters and runs over 13 times faster than BERT.
citing papers explorer
-
LegalCiteBench: Evaluating Citation Reliability in Legal Language Models
LegalCiteBench reveals that current LLMs achieve under 7% accuracy on closed-book legal citation retrieval and completion tasks, with misleading answer rates above 94% for nearly all tested models.
-
Exploring Lightweight Large Language Models for Court View Generation
Lightweight LLMs are benchmarked for court view generation and charge prediction across architectures, sizes, DNN comparisons, and task ordering on three datasets using the new CVGEvalKit framework.
-
Towards Intelligent Legal Document Analysis: CNN-Driven Classification of Case Law Texts
A 1D CNN with FastText embeddings classifies legal texts at 97.26% accuracy using 5.1 million parameters and runs over 13 times faster than BERT.