I develop open-source software for evaluating and documenting AI-generated data in political and social science research. Package documentation is built and hosted separately so that it tracks the code on every release.
craft: A CRAFT Pipeline for Evaluating LLM-Generated Data
Yuehong Cassandra Tai and Hyein Ko · R · v0.1.0 · Active development
An R package implementing the five-step C-R-A-F-T framework for evaluating data generated by large language models. It pairs reliability with validity metrics in a single workflow and propagates measurement uncertainty into downstream inference.