LLM-Driven NPCs

Complete framework to run LLM-driven NPC evaluation framework. To see the full validator prompt, see validators.py. To see the validator failure taxonomies and scoring rubrics, see dimensions.py. To see system prompts and NPC prompt, see evaluation.py. Responses_Run_*.py and Run_*.py are the complete evaluation scores for each run documented in the corresponding literature. Run world_creation.py first, then run evaluation.py. Both files give instructions as to how to run. world_creation.py will give instructions through the terminal. evaluation.py gives instructions at the top of the file.