ZKPROV: A Zero-Knowledge Approach to Dataset Provenance for Large Language Models

TL;DR

ZKPROV introduces a cryptographic framework that enables verifiable, privacy-preserving proof of dataset provenance for large language models, balancing efficiency and security in sensitive applications.

Contribution

It presents a novel zero-knowledge proof system for dataset provenance verification in LLMs, ensuring confidentiality and efficiency.

Findings

Sublinear proof generation and verification times

End-to-end overhead under 3.3 seconds for 8B parameter models

Formal security guarantees for dataset confidentiality

Abstract

As large language models (LLMs) are used in sensitive fields, accurately verifying their computational provenance without disclosing their training datasets poses a significant challenge, particularly in regulated sectors such as healthcare, which have strict requirements for dataset use. Traditional approaches either incur substantial computational cost to fully verify the entire training process or leak unauthorized information to the verifier. Therefore, we introduce ZKPROV, a novel cryptographic framework allowing users to verify that the LLM's responses to their prompts are trained on datasets certified by the authorities that own them. Additionally, it ensures that the dataset's content is relevant to the users' queries without revealing sensitive information about the datasets or the model parameters. ZKPROV offers a unique balance between privacy and efficiency by binding training datasets, model parameters, and responses, while also attaching zero-knowledge proofs to the responses generated by the LLM to validate these claims. Our experimental results demonstrate sublinear scaling for generating and verifying these proofs, with end-to-end overhead under 3.3 seconds for models up to 8B parameters, presenting a practical solution for real-world applications. We also provide formal security guarantees, proving that our approach preserves dataset confidentiality while ensuring trustworthy dataset provenance.