All Source Checks
partial85% confidence
1 evidence check
Last checked: 4/3/2026
The source does not mention the size of the interpretability team. The source does not mention that Anthropic's interpretability team is among the largest concentrations globally focused on this research agenda.
Evidence — 1 source, 1 check
partial85%Haiku 4.5 · 4/3/2026
Found: Anthropic uses dictionary learning to identify and monitor millions of neural features, mapping them to human-interpretable concepts. The interpretability team comprises an estimated <F e="anthropic" …
Note: The source does not mention the size of the interpretability team. The source does not mention that Anthropic's interpretability team is among the largest concentrations globally focused on this research agenda.
Debug info
Record type: citation
Record ID: page:anthropic:fn25