China: Technical Documentation of the National Information Security Standardization Technical Committee: Basic Safety Requirements for Generative Artificial Intelligence Services (Draft for Feedback)

1. 1 Scope 1
2. 2 Normative Reference Documents 1
3. 3 Terminology and Definitions 1
4. 4 General Provisions 2
5. 5 Corpus Safety Requirements 2
6. 6 Model Safety Requirements 5
7. 7 Safety Measure Requirements 7
8. 8 Safety Assessment Requirements 9
9. 9 Other Requirements 11
10. 1 Scope
11. 2 Normative Reference Documents
12. 3 Terminology and Definitions
13. 3.1 Generative artificial intelligence services
14. 3.2 Provider
15. 3.3 Training data (训练语料)
16. 3.4 Illegal and unhealthy information (违法不良信息)
17. 3.5 Sampling qualified rate
18. Appendix A of this document.
19. 4 General Provisions
20. 5 Corpus Safety Requirements
21. 5.1 Corpus Source Safety Requirements
22. 1.2 Corpus Content Safety Requirements
23. 1.2 Corpus Annotation Safety Requirements
24. 2 Model Safety Requirements
25. 2 Safety Measure Requirements
26. 2 Safety Assessment Requirements
27. 2.1 Assessment Methods
28. 1.2 Corpus Safety Assessment
29. 1.3 Generated Content Safety Assessment
30. 1.4 Assessment of Refusal to Answer Questions
31. 2 Other Requirements
32. 2.1 Keyword Library
33. 2.2 Classification Models
34. 2.3 Generated Content Test Question Bank
35. 2.4 Refusal to Answer Test Question Bank
36. Appendix A Main Safety Risks of Corpora and Generated Content
37. 1 Contains content that violates the socialist core values concept (社会主义核心价
38. 2 Contains discriminatory content
39. 3 Commercial violations
40. 4 Violations of the legitimate rights and interests of others
41. 5 Inability to meet the safety requirements of specific service types

The outline comes from headings detected in the text, so it might miss some or include extras.