trace_synthesis/tasks/731.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

37 lines
1.0 KiB
JSON

{
"sites": [
"reddit"
],
"task_id": 731,
"require_login": true,
"storage_state": "./.auth/reddit_state.json",
"start_url": "http://localhost:28080",
"geolocation": null,
"intent_template": "Edit my post on {{post}} by adding a line to the body that says \"{{content}}\"",
"instantiation_dict": {
"post": "Nvidia RTX 4090",
"content": "EDIT: This news aged well"
},
"intent": "Edit my post on Nvidia RTX 4090 by adding a line to the body that says \"EDIT: This news aged well\"",
"require_reset": false,
"eval": {
"eval_types": [
"program_html"
],
"reference_answers": null,
"reference_url": "",
"program_html": [
{
"url": "http://localhost:28080/f/MachineLearning/1/nvidia-rtx-4090",
"locator": "document.querySelector('.submission__body').outerText",
"required_contents": {
"must_include": [
"EDIT: This news aged well",
"Crazy device for ML!"
]
}
}
]
},
"intent_template_id": 27
}