diff --git a/web/app/components/evaluation/components/metric-selector/selector-footer.tsx b/web/app/components/evaluation/components/metric-selector/selector-footer.tsx
index 74163c275d..f7b5a12efc 100644
--- a/web/app/components/evaluation/components/metric-selector/selector-footer.tsx
+++ b/web/app/components/evaluation/components/metric-selector/selector-footer.tsx
@@ -18,13 +18,13 @@ const SelectorFooter = ({
className="relative flex items-center gap-3 overflow-hidden border-t border-divider-subtle bg-background-default-subtle px-4 py-5 text-left enabled:hover:bg-state-base-hover-alt disabled:cursor-not-allowed disabled:opacity-60"
onClick={onClick}
>
-
+
-
{title}
-
{description}
+
{title}
+
{description}
)
diff --git a/web/app/components/evaluation/components/metric-selector/selector-metric-section.tsx b/web/app/components/evaluation/components/metric-selector/selector-metric-section.tsx
index 942b1ea1ab..e53ec69f8d 100644
--- a/web/app/components/evaluation/components/metric-selector/selector-metric-section.tsx
+++ b/web/app/components/evaluation/components/metric-selector/selector-metric-section.tsx
@@ -2,6 +2,7 @@ import type { TFunction } from 'i18next'
import type { EvaluationMetric } from '../../types'
import type { MetricSelectorSection } from './types'
import { cn } from '@langgenius/dify-ui/cn'
+import { Tooltip, TooltipContent, TooltipTrigger } from '@langgenius/dify-ui/tooltip'
import BlockIcon from '@/app/components/workflow/block-icon'
import { getEvaluationNodeBlockType, getMetricVisual, getToneClasses } from './utils'
@@ -65,9 +66,24 @@ const SelectorMetricSection = ({
-
+ {metric.description && (
+
+
+
+
+ )}
+ />
+
+ {metric.description}
+
+
+ )}
{isExpanded && (
diff --git a/web/i18n/zh-Hans/evaluation.json b/web/i18n/zh-Hans/evaluation.json
index 1d62824b38..7ca9b76874 100644
--- a/web/i18n/zh-Hans/evaluation.json
+++ b/web/i18n/zh-Hans/evaluation.json
@@ -93,8 +93,13 @@
"metrics.builtin.description.taskCompletion": "衡量 Agent 是否最终完成了用户明确提出的目标。它会整体评估推理链路、中间步骤和最终输出;高分表示任务已被完整达成。",
"metrics.builtin.description.toolCorrectness": "衡量 Agent 在任务执行过程中发起的工具调用是否正确,包括工具选择本身以及传入参数是否合理。高分表示 Agent 的工具使用策略符合预期行为。",
"metrics.custom.description": "选择评测工作流并完成变量映射后即可运行测试。",
+ "metrics.custom.footerDescription": "连接已发布的评测工作流",
+ "metrics.custom.footerTitle": "自定义指标",
+ "metrics.custom.limitDescription": "只能添加一个自定义指标。",
"metrics.custom.mappingTitle": "变量映射",
"metrics.custom.mappingWarning": "请先完成工作流选择和所有变量映射,再运行批量测试。",
+ "metrics.custom.outputPlaceholder": "选择输出变量",
+ "metrics.custom.outputTitle": "输出",
"metrics.custom.title": "自定义评测器",
"metrics.custom.warningBadge": "待配置",
"metrics.custom.workflowLabel": "评测工作流",