Core Concepts
Anthropic introduces Claude 2, an upgraded AI chatbot with enhanced capabilities and improved performance, catering to user needs for longer responses and better analysis.
Abstract
Anthropic has launched Claude 2, a more advanced version of its AI chatbot, promising longer answers, detailed reasoning, fewer hallucinations, and improved performance. With an increased token capacity of up to 100,000 tokens compared to the previous version's limit of 9,000 tokens, Claude 2 can now handle more extensive data processing tasks. Additionally, Claude 2 boasts coding improvements and a higher score on the Codex HumanEval benchmark. The AI system is guided by foundational principles of fairness and autonomy and has undergone rigorous testing to enhance emotional stability and response accuracy.
Stats
Claude 2 scores in the 90th percentile of graduate school applicants on GRE reading and writing exams.
Token capacity increased to handle up to 100,000 tokens.
Coding capabilities improved with a score of 71.2% on the Codex HumanEval benchmark.
Quotes
"Compared to its predecessor Claude 1.3, Claude 2 is reportedly '2x better at giving harmless responses.'"
"If you’re already subscribed to the Claude 1.3 API, great news, you’ll be automatically rolled over to Claude 2 at no extra charge."