Directory
30 groups studying what goes on inside AI models.
Who reads models from the inside, and who studies whether there is anyone in there. Companies, universities and nonprofits, with what each one focuses on.
Reading models from the inside (interpretability)
| Name | Type | Where | Focus |
|---|---|---|---|
| Anthropic interpretability team ↗ | Company | San Francisco | Features, circuits and attribution graphs in Claude; publishes at Transformer Circuits. |
| Google DeepMind mechanistic interpretability team ↗ | Company | London | Sparse autoencoders and open tools for Gemma models. |
| OpenAI interpretability ↗ | Company | San Francisco | Sparse autoencoders and automated explanation of neurons. |
| Goodfire ↗ | Company | San Francisco | Founded 2024. Interpretability tools for auditing and editing models (Ember). |
| Apollo Research ↗ | Nonprofit | London | Interpretability and evaluations aimed at detecting deception. |
| Transluce ↗ | Nonprofit | San Francisco | Open tools for understanding model behavior at scale. |
| Decode Research (Neuronpedia) ↗ | Nonprofit | Remote | An open platform for browsing features found in models. |
| EleutherAI ↗ | Nonprofit | Remote | Open-source interpretability research and tooling. |
| FAR AI ↗ | Nonprofit | Berkeley | Safety research including interpretability. |
| Redwood Research ↗ | Nonprofit | Berkeley | Causal methods for testing explanations of model behavior. |
| Leap Laboratories ↗ | Company | London | Model-agnostic interpretability. |
| Bau Lab and the National Deep Inference Fabric ↗ | University | Northeastern University, Boston | Locating and editing knowledge inside models; shared infrastructure for studying large models. |
| Timaeus ↗ | Nonprofit | Remote | Developmental interpretability: how structure forms during training. |
| Guide Labs ↗ | Company | San Francisco | Models designed to be interpretable from the start. |
| MATS ↗ | Training program | Berkeley and London | Trains new interpretability researchers under established mentors. |
Consciousness and welfare of AI systems
| Name | Type | Where | Focus |
|---|---|---|---|
| Eleos AI Research ↗ | Nonprofit | United States | AI welfare and moral patienthood; ran the external welfare assessment of Claude 4 and the Eleos conference. |
| NYU Center for Mind, Ethics, and Policy | University | New York | Moral status of nonhuman minds, including AI; co-authors of Taking AI Welfare Seriously. |
| Anthropic model welfare program ↗ | Company | San Francisco | The first formal model welfare program at a major AI company, started 2025. |
| Google DeepMind | Company | London | Reported to employ researchers on machine consciousness and AI welfare. |
| Meta | Company | Menlo Park | Reported in 2026 to be expanding research into machine consciousness and AI welfare. |
| Reciprocal Research | Nonprofit | Named alongside Eleos as a dedicated organization in this field. | |
| PRISM | Nonprofit | Launched 2025; research into sentient machines. | |
| California Institute for Machine Consciousness ↗ | Nonprofit | California | Launched 2025; building and testing theories of machine consciousness. |
| Sussex Centre for Consciousness Science ↗ | University | Brighton, UK | Scientific study of consciousness; a leading skeptical voice on machine consciousness. |
| Mila ↗ | University | Montreal | Home of several authors of the 2023 indicator report. |
| Conscium ↗ | Company | London | Research on machine consciousness and verifying AI agents. |
| Araya ↗ | Company | Tokyo | Consciousness research and its application to AI. |
| Sentience Institute ↗ | Nonprofit | United States | Surveys and social science on attitudes to digital minds. |
| Rethink Priorities ↗ | Nonprofit | United States | Research on the probability and moral weight of digital minds. |
| Association for Mathematical Consciousness Science ↗ | Society | International | Formal and mathematical approaches to consciousness. |
About this list
Last checked 2026-11-17. Missing a group, or something wrong? Email hello@aiassistantlabs.com.