trace_synthesis/tasks/384.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
840 B
JSON

{
"sites": [
"shopping"
],
"task_id": 384,
"require_login": true,
"storage_state": "./.auth/shopping_state.json",
"start_url": "http://localhost:28082",
"geolocation": null,
"intent_template": "List the customer names who complain about the quality of EYZUTAK phone cases",
"instantiation_dict": {},
"intent": "List the customer names who complain about the quality of EYZUTAK phone cases",
"require_reset": false,
"eval": {
"eval_types": [
"string_match"
],
"reference_answers": {
"must_include": [
"Lisa Lee",
"Evelyn Kurver",
"Amanda",
"N Randall"
]
},
"reference_url": "",
"program_html": [],
"string_note": "",
"reference_answer_raw_annotation": "Lisa Lee, Evelyn Kurver, Amanda, N Randall"
},
"intent_template_id": 666
}