trace_synthesis/tasks/545.json
yuyr a84d51a101 1. 增加r1生成综合策略代码和输出;
2. 增加tasks;
3. 增加analysis部分,对策略进行归纳分类,然后进行评测。
2025-04-17 17:40:15 +08:00

40 lines
1.5 KiB
JSON

{
"sites": [
"shopping_admin"
],
"task_id": 545,
"require_login": true,
"storage_state": "./.auth/shopping_admin_state.json",
"start_url": "http://localhost:28083/admin",
"geolocation": null,
"intent_template": "Update the description of {{product}} to highlight the real user positive reviews by quoting the comments",
"instantiation_dict": {
"product": "Radiant Tee"
},
"intent": "Update the description of Radiant Tee to highlight the real user positive reviews by quoting the comments",
"require_reset": false,
"eval": {
"eval_types": [
"program_html"
],
"reference_answers": null,
"reference_url": "",
"program_html": [
{
"url": "http://localhost:28083/admin/../radiant-tee.html",
"locator": "document.querySelector('.data.item.content').outerText + (document.querySelector('.product.attribute.overview [itemprop=\"description\"]')?.outerText || '')",
"required_contents": {
"must_include": [
"What I rally love here is that it does the job of keeping me cool and dry",
"I'm a big guy and sweat A LOT",
"Even after a day of gulf, I'm still dry and comfortable",
"What a versatile shirt",
"Not only does it feel very soft compared to my old worn out polos, but it also does the job promised",
"I like going out after my game for drinks so I look good then too and don't need to change into something fresh"
]
}
}
]
},
"intent_template_id": 251
}