trace_synthesis/tasks/187.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
816 B
JSON

{
"sites": [
"shopping_admin"
],
"task_id": 187,
"require_login": true,
"storage_state": "./.auth/shopping_admin_state.json",
"start_url": "http://localhost:28083/admin",
"geolocation": null,
"intent_template": "Give me the {{Attribute}} of the products that have {{N}} units left",
"instantiation_dict": {
"Attribute": "SKU",
"N": "1-3"
},
"intent": "Give me the SKU of the products that have 1-3 units left",
"require_reset": false,
"eval": {
"eval_types": [
"string_match"
],
"reference_answers": {
"must_include": [
"WH11-S-Blue",
"WS08-XS-Blue"
]
},
"reference_url": "",
"program_html": [],
"string_note": "",
"reference_answer_raw_annotation": "WH11-S-Blue, WS08-XS-Blue"
},
"intent_template_id": 368
}