Terminal X's evaluation agent, which uses GPT-6.1, achieved a match accuracy of 88.6% for senior investor judgments, identifying deviations from fund mandates 25% more effectively than the previous solution. The system evaluates outputs based on a 54-criterion rubric.
מקור: OpenAI News — לכתבה המלאה
כתיבת תגובה