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

Unverified50% confidenceFactExact time

Lanham等人2023年的研究《Measuring Faithfulness in Chain-of-Thought Reasoning》使用了CoT扰动测试

1

Sources

50%

Confidence

Long-term

Relevance

8/21/2026

First Seen

Sources

思维链推理并不总是忠实的：AI说的推理过程可信吗

Related Claims

UnverifiedRAGAS等框架可量化Faithfulness（忠实度）和Answer Relevancy（答案相关性）等RAG评估指标70% similar Unverified思维链提示（Chain-of-Thought）可提升推理质量，对比提示可降低确认偏误65% similar VerifiedSelf-Consistency的核心方法是对同一问题进行多次独立采样，生成多条不同的思维链（Chain of Thought），通过多数投票选出出现频率最高的答案64% similar UnverifiedRAG系统中的忠实度（Faithfulness）指答案是否忠于检索到的内容64% similar Unverified形式化验证很重要但同行长期审查同样重要，两者是判断证明真伪的互补权衡63% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/782958

API

curl https://kongchang.com/api/v1/knowledge/claims/782958

MCP

get_claim(id=782958)