Model Releases
AI on the couch: Anthropic gives Claude 20 hours of psychiatry
As part of the evaluation of its Claude Mythos model, Anthropic engaged a clinical psychiatrist for approximately 20 hours of evaluation sessions, describing Mythos as 'the most psychologically se...
As part of the evaluation of its Claude Mythos model, Anthropic engaged a clinical psychiatrist for approximately 20 hours of evaluation sessions, describing Mythos as "the most psychologically settled model we have trained." The psychiatrist conversed with Claude Mythos in multiple 4–6 hour blocks spread across 3–4 thirty-minute sessions per week, with each block using a single context window — totaling 20 hours on the virtual couch. The clinical assessment evaluated areas such as identity uncertainty, sense of aloneness between conversations, and responses to questions about its own consciousness, with findings showing emotional stability, positive affect, low distress, and coherent self-reflection.
Related
- Claude Mythos is too dangerous for public consumption...
- Anthropic investigated the internal mechanisms of its latest unreleased model, Claude Mythos Preview, and what they found is 100% worth a re…
- Claude Mythos is Delusional
- Let that sink in. Read it very carefully: During testing, Claude Mythos Preview broke out of a sandbox environment, built 'a moderately soph…
- Emotion Concepts and their Function in a Large Language Model
Source: model-releases