trace_synthesis/tasks/203.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

33 lines
880 B
JSON

{
"sites": [
"shopping_admin"
],
"task_id": 203,
"require_login": true,
"storage_state": "./.auth/shopping_admin_state.json",
"start_url": "http://localhost:28083/admin",
"geolocation": null,
"intent_template": "Get the {{attribute}} of the {{status}} order",
"instantiation_dict": {
"attribute": "purchase date and order id",
"status": "most recent pending"
},
"intent": "Get the purchase date and order id of the most recent pending order",
"require_reset": false,
"eval": {
"eval_types": [
"string_match"
],
"reference_answers": {
"fuzzy_match": [
"order id: 000000299",
"purchase date: May 31, 2023"
]
},
"reference_url": "",
"program_html": [],
"string_note": "",
"reference_answer_raw_annotation": "000000299, May 31, 2023, 2:55:09 AM"
},
"intent_template_id": 366
}