Merkle-enhanced training data audit process

A computer program product (CPP) for verifying data using Merkle proof and root, the CPP comprising:

• one or more computer-readable storage media and program instructions stored on the one or more computer-readable storage media capable of performing a method, the method comprising:
• organizing a dataset into a Merkle tree, wherein the dataset includes a plurality of data elements;
• training a large language model (LLM) using the dataset;
• generating a result from a prompt used in the trained LLM;
• generating a Merkle root that represents the dataset organized into the Merkle tree;
• providing a generated Merkle proof for a specific data element of the plurality of data elements;
• fine-tuning the trained LLM with an additional data element added to the dataset after training the LLM and generating the Merkle root;
• identifying changes in the Merkle tree based on the additional data element;
• storing the Merkle root in metadata of the LLM;
• and
• verifying that the result is informed by the specific data element and the additional data element by determining whether the specific data element and the additional data element are included in the dataset using the Merkle root and the Merkle proof.

A computer system (CS) for verifying data using Merkle proof and root, the CS comprising:

• one or more computer processors, one or more computer-readable storage media, and program instructions stored on the one or more computer-readable storage media for execution by at least one of the one or more computer processors capable of performing a method, the method comprising:
• organizing a dataset into a Merkle tree, wherein the dataset includes a plurality of data elements;
• training a large language model (LLM) using the dataset;
• generating a result from a prompt used in the trained LLM;
• generating a Merkle root that represents the dataset organized into the Merkle tree;
• providing a generated Merkle proof for a specific data element of the plurality of data elements;
• fine-tuning the trained LLM with an additional data element added to the dataset after training the LLM and generating the Merkle root;
• identifying changes in the Merkle tree based on the additional data element;
• storing the Merkle root in metadata of the LLM;
• and
• verifying that the result is informed by the specific data element and the additional data element by determining whether the specific data element and the additional data element are included in the dataset using the Merkle root and the Merkle proof.