Unverified60% confidenceFactExact time
LlamaGuard 2 introduced in 2024 13 risk categories aligned with the MLCommons AI Safety Benchmark and supports holistic judgment across multi-turn conversation context.
2
Sources
60%
Confidence
Long-term
Relevance
7/9/2026
First Seen
Sources
Fable 5对决GPT-5.6 Sol：基准测试、定价与AI权力博弈深度解析
bilibili黑纹白斑马7/3/2026
Related Claims
VerifiedThe prompt injection security threat category was first systematically proposed by security researcher Simon Willison in 2022.67% similarUnverifiedThe DevFest 2026 theme focuses on building, securing, and scaling in the age of Agentic AI.63% similarUnverifiedMeta open-sourced the Llama Guard series at the end of 2023; it performs safety classification based on large language models, supports custom taxonomies, and currently remains primarily text-based.63% similarUnverifiedMITRE在2023年发布了面向AI/ML系统的威胁知识库MITRE ATLAS62% similarVerifiedOWASP has listed "insecure AI-generated code" as one of the key risks to focus on in 2025.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/339084API
curl https://kongchang.com/api/v1/knowledge/claims/339084MCP
get_claim(id=339084)