Big model execution method and system based on ownership retrieval enhancement and medium

The invention relates to the technical field of ownership intelligence, and particularly discloses a big model execution method and system based on ownership retrieval enhancement and a medium, and the execution method mainly comprises the steps: obtaining an execution example of an ownership agent, the execution example being a task instruction I and an action sequence A based on the task instruction I; constructing a retrieval library C which takes the task instruction I as a key and takes the action sequence A based on the task instruction I as a value; calculating a correlation score (Score); and constructing an execution logic set # imgabs0 # by using the task instruction I and the action sequence A according to the correlation score Score, and constructing output content according to the execution logic set # imgabs1 #. According to the method, related execution examples are adaptively retrieved by obtaining the execution examples, then the examples serve as a part of the context, key information is ext