1 link tagged with all of: language-models + model-transparency + introspection + concept-injection + neural-activations
Links
Researchers used a “concept injection” method to compare Claude’s self-reported thoughts with its actual neural activity. They found Claude Opus 4 and 4.1 sometimes detect and control injected concepts, suggesting limited but real introspective abilities that improve with model capacity.
introspection
concept-injection
neural-activations
language-models
model-transparency