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

条件概率的数学表达为P(A|B)=P(A∩B)/P(B)，指在已知某一事件发生的前提下另一事件发生的概率

1

Sources

50%

Confidence

Long-term

Relevance

7/20/2026

First Seen

Sources

GPT-5.6 Codex三模型实测：Sol/Terra/Luna糖果逻辑题横评

Related Claims

Unverified策略梯度定理证明梯度可表示为∇J(θ) = E[∇log π(a|s;θ) · Q^π(s,a)]，使用log-derivative trick将对概率分布的求导转化为对log概率的求导乘以采样值65% similar Unverified逻辑回归通过Sigmoid函数将线性组合的输出压缩到(0,1)区间，表示样本属于某一类别的概率63% similar Unverified自回归语言模型的串行约束源于其概率分解形式P(x₁,...,xₙ)=∏P(xₜ|x₁,...,xₜ₋₁)，是内在数学属性而非算法缺陷63% similar UnverifiedPrompt本质上是模型的'条件'，条件越精确，模型的概率分布就越集中在用户期望的输出上，而不是分散在各个方向。62% similar Unverified引导式生成模型提供的概率数字本质上是'受过良好约束的推断'而非可直接核实的真相62% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/573318

API

curl https://kongchang.com/api/v1/knowledge/claims/573318

MCP

get_claim(id=573318)