trace_synthesis/tasks/464.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

37 lines
1.3 KiB
JSON

{
"sites": [
"shopping_admin"
],
"task_id": 464,
"require_login": true,
"storage_state": "./.auth/shopping_admin_state.json",
"start_url": "http://localhost:28083/admin",
"geolocation": null,
"intent_template": "Update the product description of {{product}} to highlight the real user positive reviews by quoting the comments",
"instantiation_dict": {
"product": "Antonia Racer Tank"
},
"intent": "Update the product description of Antonia Racer Tank 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/../antonia-racer-tank.html",
"locator": "document.querySelector('.data.item.content').outerText + (document.querySelector('.product.attribute.overview [itemprop=\"description\"]')?.outerText || '')",
"required_contents": {
"must_include": [
"This is in regular rotation at the gym",
"Its colorful and looks kinda cute under my exercise tanks",
"it's very stylish for yoga or something else low impact"
]
}
}
]
},
"intent_template_id": 251
}