PH-LegalBench: Measuring Legal Reasoning on Philippine Laws
Adapting LegalBench to Philippine law: collecting case law and building tasks that measure LLM legal reasoning.
About the project
PH-LegalBench is an exciting opportunity to adapt the groundbreaking LegalBench project for the Philippine legal system. While LegalBench established a comprehensive evaluation suite for legal reasoning across multiple jurisdictions, this Philippine adaptation addresses the unique characteristics, complexities, and linguistic requirements of Philippine legislation. This project focuses on developing legal reasoning tests that reflect real-world Philippine legal scenarios. You’ll work on creating evaluation tasks that assess AI’s ability to interpret Philippine statutes, analyze local court decisions, and determine compliance with national regulations.
As a mentee, you’ll contribute to building essential tools that will help advance legal technology in the Philippines, making legal services more accessible and accurate for everyone. This is your chance to work on cutting-edge AI research with direct impact on the Philippine justice system.
What mentees potentially work on
- mostly data work including automating collection of legal cases from online Philippine case law repositories, preprocessing the collected data to make sure it’s “benchmark-ready”
- prototyping evaluation pipeline to test LLMs’ legal reasoning using the collected benchmark data
Requirements and recommended background
- experience with evaluating LLMs or AI agents
- experience with building or collecting datasets online
- interest with Philippine laws
There are no formal eligibility limits. See who can apply; you can still apply if you don't match every item.