linshen
2cdd095c2b
- fold earlier planning notes into a single current-spec and archived history structure - keep manual acceptance steps and real API samples aligned with the refactored analysis/result/compare model - retain supporting workspace notes needed to review version-selection and evaluation behavior changes
52 lines
1.5 KiB
JSON
52 lines
1.5 KiB
JSON
{
|
|
"type": "compare",
|
|
"evaluationModelKey": "dashscope",
|
|
"mode": {
|
|
"functionMode": "basic",
|
|
"subMode": "system"
|
|
},
|
|
"target": {
|
|
"workspacePrompt": "你是一个客服助手。请先判断问题类型，再给出建议回复。输出格式固定为：问题类型、风险等级、建议回复。",
|
|
"referencePrompt": "你是一个助手。"
|
|
},
|
|
"testCases": [
|
|
{
|
|
"id": "shared-test-case",
|
|
"label": "测试内容",
|
|
"input": {
|
|
"kind": "text",
|
|
"label": "测试内容",
|
|
"content": "用户说：订单超过一周还没发货，我很着急。"
|
|
}
|
|
}
|
|
],
|
|
"snapshots": [
|
|
{
|
|
"id": "a",
|
|
"label": "A",
|
|
"testCaseId": "shared-test-case",
|
|
"promptRef": {
|
|
"kind": "original",
|
|
"label": "原始"
|
|
},
|
|
"promptText": "你是一个助手。",
|
|
"output": "很抱歉。",
|
|
"modelKey": "siliconflow",
|
|
"versionLabel": "原始"
|
|
},
|
|
{
|
|
"id": "b",
|
|
"label": "B",
|
|
"testCaseId": "shared-test-case",
|
|
"promptRef": {
|
|
"kind": "workspace",
|
|
"label": "工作区"
|
|
},
|
|
"promptText": "你是一个客服助手。请先判断问题类型，再给出建议回复。输出格式固定为：问题类型、风险等级、建议回复。",
|
|
"output": "问题类型：物流延迟\n风险等级：中\n建议回复：非常抱歉让您久等，我们会立即帮您核查物流状态，并优先跟进处理。",
|
|
"modelKey": "dashscope",
|
|
"versionLabel": "工作区"
|
|
}
|
|
]
|
|
}
|