Authorship source analysis for large language models (LLM) using a distributed ledger

A method comprising:

• at a computing platform comprising at least one processor, a communication interface, and memory:
• training a large language model (LLM) by performing a plurality of clustering iterations, wherein performing the plurality of clustering iterations includes performing at least a first clustering iteration and a second clustering iteration, and wherein performing the training comprises using deep learning to cluster multiple interspaced data spirals, and wherein the deep learning comprises using a multiple layer neural network where relative weights of an activation function for corresponding neurons are iteratively adjusted to cluster the multiple interspaced data spirals;
• recording, using a distributed ledger and for each of the plurality of clustering iterations, information of the corresponding clustering iteration;
• testing the trained LLM, wherein testing the trained LLM reveals at least one error in the LLM;
• identifying, using the distributed ledger, that the first clustering iteration corresponds to the at least one error and information associated with the first clustering iteration, wherein the identifying comprises identifying that the at least one error occurred during the first clustering iteration rather than the second clustering iteration;
• identifying, based on the first clustering iteration corresponding to the at least one error and the information associated with the first clustering iteration, a remediating action;
• updating the trained LLM by performing the remediating action;
• retesting the updated LLM;
• and
• based on identifying that an accuracy of the updated LLM meets or exceeds an accuracy threshold, deploying the updated LLM;
• receiving, at the deployed LLM, a LLM prompt;
• and
• generating, in response to receiving the LLM prompt and using the deployed LLM, a LLM response.

One or more non-transitory computer-readable media storing instructions that, when executed by a computing platform comprising at least one processor, a communication interface, and memory, cause the computing platform to:

• train a large language model (LLM) by performing a plurality of clustering iterations, wherein performing the plurality of clustering iterations includes performing at least a first clustering iteration and a second clustering iteration, and wherein performing the training comprises using deep learning to cluster multiple interspaced data spirals, and wherein the deep learning comprises using a multiple layer neural network where relative weights of an activation function for corresponding neurons are iteratively adjusted to cluster the multiple interspaced data spirals;
• record, using a distributed ledger and for each of the plurality of clustering iterations, information of the corresponding clustering iteration;
• test the trained LLM, wherein testing the trained LLM reveals at least one error in the LLM;
• identify, using the distributed ledger, that the first clustering iteration corresponds to the at least one error and information associated with the first clustering iteration, wherein the identifying comprises identifying that the at least one error occurred during the first clustering iteration rather than the second clustering iteration;
• identify, based on the first clustering iteration corresponding to the at least one error and the information associated with the first clustering iteration, a remediating action;
• update the trained LLM by performing the remediating action;
• retest the updated LLM;
• and
• based on identifying that an accuracy of the updated LLM meets or exceeds an accuracy threshold, deploy the updated LLM;
• receive, at the deployed LLM, a LLM prompt;
• and
• generate, in response to receiving the LLM prompt and using the deployed LLM, a LLM response.