Silico
Interpretability agent that explains, debugs and precisely controls AI model behavior
Community-submitted · content unverified. This organization has not verified ownership or the information in this profile.
About
Silico is Goodfire's interpretability agent. Given a research goal it plans experiments, runs them in parallel across nodes and returns inspectable results, using skills and libraries for auto-interpretation, probes and sparse autoencoders, model diffing, causal analysis and data attribution. Teams use it to explain what a model has learned, trace regressions and unexpected behaviors, and deploy guardrails against unsafe agent behavior, with versions for LLMs, life sciences and robotics and vision.
- Website
- goodfire.com/silico
- Added via
- web
- Ownership
- unclaimed
Where to go
For people
- Homepagegoodfire.com/silico
For agents
Nothing listed yet.
Product details
- Product kind
- Saas