Highlights from the Claude 4 system prompt
SMRTR summary
Anthropic has released system prompts for their Claude Opus 4 and Claude Sonnet 4 models, providing insight into how these AI assistants function. The prompts establish Claude's personality, knowledge cutoff date, safety guidelines, and interaction style. They also include instructions for using tools like web search and code generation. While Anthropic published most of the prompts, some details about tools were leaked separately. This transparency offers valuable information for users to better understand and utilize Claude's capabilities.
SMRTR provides this summary for quick context. The original article belongs to Daily.dev.
Read the original article