awesome-ai-alignment
awesome-ai-alignment copied to clipboard
Add MirrorDNA - reflective AI alignment framework
MirrorDNA
MirrorDNA is a practical framework for human-AI alignment through runtime protocol rather than training-time approaches.
Key Features
- Three Laws: Truth (Fact/Estimate/Unknown), Vault (sovereign memory), Mesh (multi-model coordination)
- MESH BOOT: Universal alignment prompt for any LLM
- Open source: MIT licensed, actively maintained
Links
- Repository: https://github.com/MirrorDNA-Reflection-Protocol/active-mirror-identity
- Paper: https://doi.org/10.5281/zenodo.17787619
- Creator: Paul Desai (@pauldesai123)
Summary by Sourcery
Documentation:
- Add a README section describing the MirrorDNA runtime alignment framework, its key concepts, and links to the upstream project resources.