trace_synthesis/tasks/57.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

38 lines
1015 B
JSON

{
"sites": [
"map"
],
"task_id": 57,
"require_login": true,
"storage_state": null,
"start_url": "https://www.openstreetmap.org",
"geolocation": null,
"intent_template": "Tell me the closest {{place1}}(s) to {{place2}}",
"instantiation_dict": {
"place1": "restaurant",
"place2": "university center at Carnegie Mellon University"
},
"intent": "Tell me the closest restaurant(s) to university center at Carnegie Mellon University",
"require_reset": false,
"eval": {
"eval_types": [
"string_match"
],
"reference_answers": {
"must_include": [
"El Gallo de Oro",
"Back Bar Grill",
"Grano",
"Beefsteak",
"Nourish",
"Schatz Dining Room",
"Au Bon Pain"
]
},
"reference_url": "",
"program_html": [],
"string_note": "",
"reference_answer_raw_annotation": "El Gallo de Oro, Back Bar Grill, Grano, Beefsteak, Nourish, Schatz Dining Room, Au Bon Pain"
},
"intent_template_id": 69
}