Unverified50% confidenceFactExact time
The industry typically uses SLO (Service Level Objective) and Error Budget to quantify impacts such as performance degradation, for example using '99.9% of requests return within 500ms' as a target.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe inference cost of SLMs can be 1 to 2 orders of magnitude lower than mainstream large models, and they can run locally or on edge devices.63% similarUnverified工业SCADA故障事件数据高度不平衡，故障事件通常占比不足1%，普通分类模型会偏向正常状态导致低召回率62% similarUnverified流失率（Churn Rate）最初是SaaS行业中衡量客户流失速度的核心概念，表示特定时间段内取消订阅的用户比例61% similarUnverified研究表明，即使在 GSM8K 或 MATH 数据集上接近满分的模型，面对稍作变形或需要跨领域组合的问题时性能下降可达30-50个百分点61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1031346API
curl https://kongchang.com/api/v1/knowledge/claims/1031346MCP
get_claim(id=1031346)