Internal evaluation benchmark for testing long form multi-character script generation.
Which model writes My Weird Prompts, and how others scored