Unverified85% confidenceFactTime unknown
Petri是Anthropic开发的一款开源工具，专门用于测试和评估AI模型的对齐表现
1
Sources
85%
Confidence
Long-term
Relevance
5/31/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPetri在AI安全技术体系中属于评估环节，不负责训练模型，而是在模型训练完成后系统性地检验对齐效果80% similarUnverifiedVeta is an open-source AI testing Agent designed specifically for the Android platform that, given a test objective described in natural language, can autonomously plan, execute, verify, and generate reports.70% similarUnverifiedVeta generates a complete action trace for each test run, containing screenshots, LLM reasoning output, executed action instructions, and assertion verification results.59% similarUnverifiedOpenAI introduced monitoring upgrades such as activation probes as an engineering application of mechanistic interpretability research to detect potentially harmful intent.59% similarUnverified朱雀检测工具采用类似生成对抗网络（GAN）判别器的架构，通过人工标注的真人写作与AI生成样本进行有监督训练57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/12272API
curl https://kongchang.com/api/v1/knowledge/claims/12272MCP
get_claim(id=12272)