Automated quality evaluation in compliance policy generation

A system for generating description from synthetic data used by large language (LLM), comprising:

• one or more computers with executable instructions that when executed cause the system to:
• obtain inferred code and data to determine accuracy of a description previously generated by a synthetic data stream;
• parse inferred code to determine an inferred code template and an inferred description template;
• obtain a plurality of code template sets and compare it to said inferred code template;
• determine when any of said plurality of code template sets are a close match to content of said inferred code template;
• extract a plurality of seeds from said inferred code template and any of said code template sets that were determined to be a close match to said inferred code template;
• obtain a corresponding description template corresponding to said any of said code template sets determined to be a close match to said inferred code template;
• and
• generate a new description from said synthetic data stream using said plurality of seeds and said corresponding description template.

A computer program product for generating description from synthetic data used by large language (LLM), comprising

• one or more computer readable storage media;
• program instructions, stored on at least one of the one or more storage media for execution by at least one of the one or more processors via at least one of the one or more memories, to:
• obtain inferred code and data to determine accuracy of a description previously generated by a synthetic data stream;
• parse inferred code to determine an inferred code template and an inferred description template;
• obtain a plurality of code template sets and compare it to said inferred code template;
• determine when any of said plurality of code template sets are a close match to content of said inferred code template;
• extract a plurality of seeds from said inferred code template and any of said code template sets that were determined to be a close match to said inferred code template;
• obtain a corresponding description template corresponding to said any of said code template sets determined to be a close match to said inferred code template;
• and
• generate a new description from said synthetic data stream using said plurality of seeds and said corresponding description template.