README.md

metadata

title: README
emoji: ⚖️
colorFrom: blue
colorTo: gray
sdk: static
pinned: false
license: cc-by-4.0
short_description: Benchmarks of language models in Polish law

LexAlpha Lab

LexAlpha Lab is a joint research initiative of the law firm KTZR (Gdańsk) and LexAlpha Sp. z o.o. (Warsaw). We evaluate how large language models answer questions about Polish law: correctness, hallucinated citations, and grounding in official legal sources.

• LawBenchmark.pl: https://lawbenchmark.pl
• Code and data: https://github.com/apiotrowski-afk/lawbenchmark
• Methodology working paper: https://doi.org/10.5281/zenodo.23674794

Our datasets are published for evaluation and research. Please do not use them to train or fine-tune language models.

Contact: a.piotrowski@ktzr.pl