Unverified50% confidenceBenchmarkExact time
MTEB是覆盖56项任务的综合评测基准，业界通常以其衡量嵌入模型能力
1
Sources
50%
Confidence
Long-term
Relevance
7/7/2026
First Seen
Sources
Ternlight: A 7MB Browser-Side Embedding Model for Server-Free Semantic Search
hackernewshackernews7/6/2026
Related Claims
UnverifiedMTEB is an important reference leaderboard for evaluating the performance of embedding models.79% similarUnverified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65% similarUnverified标准化基准（如MMLU、HumanEval、GSM8K）任务边界清晰、上下文适中，与真实工程调试场景存在分布偏移64% similarVerifiedMMLU evaluates a model's breadth of knowledge across 57 subject areas, HumanEval measures code generation accuracy, GSM8K tests mathematical reasoning, and HELM constructs a comprehensive multi-dimensional evaluation framework.63% similarUnverified当企业需评测数百个AI用例时，真正需要的是评估优先的能力：可复现的测试集、多轮对话评测、红队安全测试及跨团队统一评分标准63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/207855API
curl https://kongchang.com/api/v1/knowledge/claims/207855MCP
get_claim(id=207855)