Unverified50% confidenceSolutionExact time
企业级Agent评估常见维度包括准确性、相关性、忠实度（Faithfulness，RAG场景）、安全性，实际工程中将LLM自动评估与小规模人工标注结合构建持续评估流水线
1
Sources
50%
Confidence
Long-term
Relevance
7/25/2026
First Seen
Sources
Related Claims
Unverified评估指标选择需与业务目标对齐，客服Agent优先任务完成率，代码生成Agent则更关注语法正确率和可运行率71% similarUnverifiedFoundry provides built-in evaluators covering metrics such as relevance, fluency, groundedness, and Agent behavior evaluation.69% similarUnverified当前业界Agent评估主要采用轨迹评估、结果评估和效率评估三个维度的框架68% similarUnverifiedLLM evaluation needs to simultaneously consider multiple dimensions such as factual accuracy, instruction-following, harmlessness, helpfulness, and expression quality, and tensions among these dimensions are common.68% similarUnverified完整的Agent评估体系通常包含三类方式：基于规则的评估、基于参考答案的相似度评估（如BLEU、ROUGE）、以及LLM-as-Judge67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/613190API
curl https://kongchang.com/api/v1/knowledge/claims/613190MCP
get_claim(id=613190)