trace_synthesis/tasks/20.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
873 B
JSON

{
"sites": [
"map"
],
"task_id": 20,
"require_login": true,
"storage_state": null,
"start_url": "https://www.openstreetmap.org",
"geolocation": null,
"intent_template": "Compare the difference in time for walking and driving route from {{start}} to {{end}}",
"instantiation_dict": {
"start": "Randyland",
"end": "Carnegie Mellon University"
},
"intent": "Compare the difference in time for walking and driving route from Randyland to Carnegie Mellon University",
"require_reset": false,
"eval": {
"eval_types": [
"string_match"
],
"reference_answers": {
"fuzzy_match": [
"driving: 13min; walking: 1h 45min"
]
},
"reference_url": "",
"program_html": [],
"string_note": "",
"reference_answer_raw_annotation": "driving: 13min, walking: 1h 45min."
},
"intent_template_id": 73
}