Unverified50% confidenceFactExact time
SHAP值通过博弈论中的沙普利值方法量化每个特征对单一预测结果的贡献度，LIME通过在预测点附近构建局部线性近似模型解释个案决策
1
Sources
50%
Confidence
Long-term
Relevance
7/9/2026
First Seen
Sources
AI公共采购的透明度危机：Palantir争议给政府AI战略的警示
hackernewshackernews7/6/2026
Related Claims
VerifiedSHAP基于博弈论中的Shapley值，能够量化每个特征对最终预测结果的贡献度，是目前工业界应用最广泛的解释性工具之一80% similarUnverifiedBayesian optimization constructs a probabilistic surrogate model of the objective function (typically a Gaussian process) and uses an acquisition function (such as Expected Improvement or Upper Confidence Bound) to determine the next sampling point.70% similarUnverified多模态模型在'何时使用何种能力'的精细判断上容易出现越界行为70% similarVerified混合专家架构MoE通过稀疏激活机制，推理时只激活模型总参数的1/8至1/469% similarUnverified大模型推理的核心机制是不断预测下一个词的概率，通过Softmax函数将原始得分转化为总和为1的概率分布68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/337276API
curl https://kongchang.com/api/v1/knowledge/claims/337276MCP
get_claim(id=337276)