⟳
Loading news item...
Go back
Anthropic introduces introspection adapters for language models to self-report behaviors - Gloria Terminal