UlisseMini/procgen-tools

Tools for running experiments on RL agents in procgen environments

39
/ 100
Emerging

This project provides specialized tools for analyzing and understanding how reinforcement learning agents behave within the 'maze' environment of the Procgen benchmark. It allows researchers to input trained agent models and maze configurations, then observe and manipulate the agent's internal decision-making processes and resulting actions. This is primarily for AI/ML researchers studying agent interpretability, goal misgeneralization, and behavioral statistics in controlled environments.

No commits in the last 6 months.

Use this if you are an AI/ML researcher investigating the internal workings of reinforcement learning agents, specifically their decision-making in maze-like environments, and want to test hypotheses about their goal-directed behavior.

Not ideal if you are looking for a stable, production-ready library with extensive documentation or if your research involves environments other than the Procgen 'maze' or different types of AI models.

reinforcement-learning-research agent-interpretability goal-misgeneralization behavioral-analysis AI-safety-research
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 16 / 25
Community 17 / 25

How are scores calculated?

Stars

20

Forks

11

Language

Jupyter Notebook

License

MIT

Last pushed

Apr 05, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/UlisseMini/procgen-tools"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.