Anthropic Open-Sources FlashProve: Zero-Knowledge Proofs for Large Model Inference on a Single H100
Source:
www.anthropic.com
According to the Anthropic technical blog, the Anthropic Frontier Red Team has launched the FlashProve system, exploring the use of zero-knowledge succinct non-interactive arguments of knowledge to verify the reasoning processes of large language models, enabling users to confirm that responses were indeed generated by the designated model without obtaining the model's weights.