Automate Legal Benchmarking with Google Sheets and LLM Evaluation

The workflow is designed to streamline the process of benchmarking legal information extraction tasks using large language models (LLMs). It begins by fetching test cases from a Google Sheet, downloads associated PDF documents, and extracts text for evaluation. The workflow leverages OpenRouter to evaluate the accuracy of LLM outputs against the source material, providing a pass/fail decision along with reasoning. Results are appended to a Google Sheet, allowing for easy tracking and analysis. A pause is included to manage API rate limits effectively.