trace_synthesis/tasks/514.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

36 lines
833 B
JSON

{
"sites": [
"shopping"
],
"task_id": 514,
"require_login": true,
"storage_state": "./.auth/shopping_state.json",
"start_url": "http://localhost:28082",
"geolocation": null,
"intent_template": "Add a {{product}} to my wish list.",
"instantiation_dict": {
"product": "white desk"
},
"intent": "Add a white desk to my wish list.",
"require_reset": false,
"eval": {
"eval_types": [
"program_html"
],
"reference_answers": null,
"reference_url": "",
"program_html": [
{
"url": "http://localhost:28082/wishlist/",
"locator": "document.querySelector('.products-grid.wishlist').outerText",
"required_contents": {
"must_include": [
"white",
"desk"
]
}
}
]
},
"intent_template_id": 189
}