Co-initiation of a large model security assessment benchmark

The National Research Centre for the Development of Industrial Information Security, building on the technical accumulated and practical experience in the field of large-scale model security assessment, has officially launched a joint initiative to establish a large model security assessment benchmark, and has now begun to call for co-consolidation units to develop a system of artificial intelligence security assessment covering the full dimensions. The benchmark builds a panoramic assessment framework around the security governance needs of the real application of the large model, covering content security, value alignment, robustness, fairness, privacy protection, and genuine and credible six core security dimensions, and forms a systematic whole-chain assessment system。