Unverified50% confidenceFactExact time
Since the history is a sum, any linear scoring model built on it can be exactly decomposed into the independent contribution of each event: score = w·h + b = Σ_i (e_i·w) + b.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedA univariate linear regression model can be expressed as y = w * x + b, where w is the weight (slope) and b is the bias (intercept).65% similarUnverified自回归模型将联合概率拆解为逐步条件概率的乘积，通过因果注意力掩码确保每个token只能看到左侧历史信息64% similarUnverifiedIn the selective prediction framework, the utility function is typically written as U = α·P(correct) - β·P(wrong), where α and β represent the reward weight for correct answers and the penalty weight for wrong answers, respectively.62% similarUnverified"Correlation does not equal causation" is the fundamental limitation of traditional interpretability methods that infer model behavior by observing statistical associations between activation values and outputs.62% similarUnverified传统自回归推理中，模型每一步只能基于已有上下文预测下一个token，是串行依赖关系61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1028333API
curl https://kongchang.com/api/v1/knowledge/claims/1028333MCP
get_claim(id=1028333)