Unverified50% confidenceFactExact time
A Golden Eval Set is a collection of test samples pre-labeled with correct answers, used to measure AI output quality and to run regression tests after model upgrades or prompt modifications in order to detect performance degradation.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Unverified评估集构建遵循「黄金数据集」原则：从真实生产流量采样，覆盖高频场景、边界情况和已知失败案例，并由人工标注期望输出作为基准73% similarUnverifiedThe evaluation stage requires preparing a Golden Set (standard Q&A collection) covering typical scenarios in advance, running scores on both old and new rules separately for comparison, in order to prevent rule inflation.70% similarUnverified回归测试时可将历史轨迹中的工具调用序列作为黄金标准，使用编辑距离、工具调用匹配率等指标量化新版本模型或Prompt的回归风险69% similarUnverifiedThe industry typically employs three evaluation frameworks for AI model A/B testing: rule-based automated evaluation, the LLM-as-Judge method, and comparison against manually annotated golden datasets.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/906625API
curl https://kongchang.com/api/v1/knowledge/claims/906625MCP
get_claim(id=906625)