mirror of
https://github.com/linshenkx/prompt-optimizer.git
synced 2026-07-01 11:14:19 +08:00
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
39 lines
984 B
JSON
39 lines
984 B
JSON
{
|
|
"type": "result",
|
|
"score": {
|
|
"overall": 90,
|
|
"dimensions": [
|
|
{
|
|
"key": "goalAchievement",
|
|
"label": "目标达成度",
|
|
"score": 100
|
|
},
|
|
{
|
|
"key": "outputQuality",
|
|
"label": "输出质量",
|
|
"score": 95
|
|
},
|
|
{
|
|
"key": "constraintCompliance",
|
|
"label": "约束符合度",
|
|
"score": 85
|
|
},
|
|
{
|
|
"key": "promptEffectiveness",
|
|
"label": "提示词引导有效性",
|
|
"score": 90
|
|
}
|
|
]
|
|
},
|
|
"improvements": [
|
|
"可考虑在变量替换后，对'不要解释'进行更明确的语气强化（如：请仅以诗歌形式回应），以进一步确保语气稳定性。"
|
|
],
|
|
"summary": "执行结果符合古典风格诗歌创作要求且未附加额外解释，但'不要解释'的语气约束仍有微小改进空间。",
|
|
"patchPlan": [],
|
|
"metadata": {
|
|
"model": "dashscope",
|
|
"timestamp": 1773729178590,
|
|
"duration": 7759
|
|
}
|
|
}
|