Anthropic Optimizes Biological Safety Mechanism for Claude Fable 5, Reducing False Positives by 85%
Anthropic updated the biosecurity safeguards of Claude Fable 5 yesterday by adjusting the safety classifier to more accurately distinguish between harmless and high-risk queries. Official tests show an 85% reduction in degradation for biology-related Q&A, allowing users to handle a wider range of biological tasks while preventing malicious use.
Return to original text
A short introduction is retained here; the full text from the source website is not reposted. For specific wording and subsequent updates, please refer to the original article.
Open the original article on ITHome