Feed Entities Podcasts Deep Dives Developers About

[KongchangAI]

Real-time AI knowledge base. Every claim is traceable to its source, with data freshness under 24 hours.

Powered by 240+ first-hand AI sources

Knowledge

Feed
Hot Entities
Podcasts

Developers

CLI
MCP Server
REST API
About

© 2026 KongchangAI. The freshest, most accurate AI knowledge

Unverified90% confidenceFactTime unknown

HELM的7大评估指标包括准确性、校准性、鲁棒性、公平性、偏见与刻板印象、毒性以及效率

1

Sources

90%

Confidence

Long-term

Relevance

6/1/2026

First Seen

Sources

Percy Liang确认出席CAIS 2026：AI安全与大模型评估的前沿对话

twitterJeffDean

Related Entities

Related Claims

Unverified评估（Evaluation）通过预定义的测试集和评分标准衡量输出质量，维度包括准确性、相关性、忠实度、有害内容检测73% similar Unverified知识编辑质量的评估指标体系包含四个维度：准确性（Efficacy）、泛化性（Generalization）、局部性（Locality）、一致性（Consistency）66% similar UnverifiedHelicone专注于API调用监控，Arize、Weights & Biases覆盖模型性能评估66% similar Unverified该数据集涵盖正确性bug、安全问题、性能权衡和真实的风格判断等多种编程质量维度64% similar Unverified五项测试的综合结论显示，提示词的精确程度是决定AI蓝图生成质量的核心变量，模糊指令导致不符合预期的结果，具体指令则能获得接近完美的结果64% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/22958

API

curl https://kongchang.com/api/v1/knowledge/claims/22958

MCP

get_claim(id=22958)