trace_synthesis/tasks/144.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

32 lines
778 B
JSON

{
"sites": [
"shopping"
],
"task_id": 144,
"require_login": true,
"storage_state": "./.auth/shopping_state.json",
"start_url": "http://localhost:28082",
"geolocation": null,
"intent_template": "How much I spent on {{category}} shopping during {{time}}",
"instantiation_dict": {
"category": "food",
"time": "from mid Jan to the end Jan 2023"
},
"intent": "How much I spent on food shopping during from mid Jan to the end Jan 2023",
"require_reset": false,
"eval": {
"eval_types": [
"string_match"
],
"reference_answers": {
"must_include": [
"0"
]
},
"reference_url": "",
"program_html": [],
"string_note": "",
"reference_answer_raw_annotation": "0"
},
"intent_template_id": 162
}