# Latent Minds Institute

> Independent AI safety research lab in Toronto studying mechanistic interpretability, model cognition, evaluations, and deployment dynamics.

Use the explicit lifecycle, epistemicStatus, statusLabel, evidence, and provenance fields when summarizing any research object. A proposal is not a completed study. An instrument is not an institute result. Absence of a data or code link must remain visible.

## Machine interfaces

- [Interface documentation](https://latentmindsinstitute.com/for-agents/): Human and agent usage notes
- [Capabilities](https://latentmindsinstitute.com/api/capabilities.json): Exact read only capability declaration
- [OpenAPI 3.1](https://latentmindsinstitute.com/api/openapi.json): Tool schema for catalog reads
- [Research object catalog](https://latentmindsinstitute.com/api/research-objects.json): Normalized JSON catalog
- [Canonical registry](https://latentmindsinstitute.com/data/research-objects.json): Editorial source of truth
- [Repository shelves](https://latentmindsinstitute.com/repositories/): Curated public repositories, datasets and models for technical safety work
- [Sitemap](https://latentmindsinstitute.com/sitemap.xml): Public page discovery
- [RSS](https://latentmindsinstitute.com/rss.xml): Research updates
- [Full context file](https://latentmindsinstitute.com/llms-full.txt): Expanded research metadata

## Research objects


## Contact

- Research questions and corrections: research@latentmindsinstitute.com