commons-sentience-sandbox

Materials for skeptical reviewers and universities

This repository is a public research sandbox, not evidence that simulated agents are sentient and not a claim of new physics.

Start here

• Executive summary
• Research dossier
• White paper
• Continuity and current work
• Citation metadata

Review standard

Please use a falsification-first approach:

1. Reproduce the documented setup and baseline runs.
2. Inspect the source, test commands, and saved outputs.
3. Compare claims with the evidence actually produced.
4. Report failures, null results, and alternative explanations.
5. Treat simulated continuity, memory, reflection, and governance as measured software behaviors—not proof of consciousness.

Technical reviewers are especially welcome to examine reproducibility, audit logs, permission boundaries, failure handling, and whether the evaluation metrics support the stated conclusions. University reviewers may use the repository as a starting point for independent replication or a methods critique.

No special access or paid software is required for the core review path beyond Python 3.9+ and the dependencies documented in the repository.

This site is open source. Improve this page.