trace_synthesis/tasks/110.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

41 lines
1.3 KiB
JSON

{
"sites": [
"shopping_admin"
],
"task_id": 110,
"require_login": true,
"storage_state": "./.auth/shopping_admin_state.json",
"start_url": "http://localhost:28083/admin",
"geolocation": null,
"intent_template": "Presents the monthly count of successful orders {{period}} in MM:COUNT format",
"instantiation_dict": {
"period": "from Jan to Nov 2022"
},
"intent": "Presents the monthly count of successful orders from Jan to Nov 2022 in MM:COUNT format",
"require_reset": false,
"eval": {
"eval_types": [
"string_match"
],
"reference_answers": {
"fuzzy_match": [
"January: 11 orders",
"Feburary: 16 orders",
"March: 14 orders",
"April: 7 orders",
"May: 8 orders",
"June: 13 orders",
"July: 9 orders",
"August: 8 orders",
"Sepetember: 10 orders",
"Octorbor: 4 orders",
"November: 5 orders"
]
},
"reference_url": "",
"program_html": [],
"string_note": "",
"reference_answer_raw_annotation": "January: 11 orders Feburary: 16 orders March: 14 orders April: 7 orders May: 8 orders June: 13 orders July: 9 orders August: 8 orders Sepetember: 10 orders Octorbor: 4 orders November: 5 orders "
},
"intent_template_id": 270
}