trace_synthesis/tasks/23.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
968 B
JSON

{
"sites": [
"shopping"
],
"task_id": 23,
"require_login": true,
"storage_state": "./.auth/shopping_state.json",
"start_url": "http://localhost:28082/3-pack-samsung-galaxy-s6-screen-protector-nearpow-tempered-glass-screen-protector-with-9h-hardness-crystal-clear-easy-bubble-free-installation-scratch-resist.html",
"geolocation": null,
"intent_template": "List out reviewers, if exist, who mention about {{description}}",
"instantiation_dict": {
"description": "good fingerprint resistant"
},
"intent": "List out reviewers, if exist, who mention about good fingerprint resistant",
"require_reset": false,
"eval": {
"eval_types": [
"string_match"
],
"reference_answers": {
"must_include": [
"Rachel",
"T. Gannon"
]
},
"reference_url": "",
"program_html": [],
"string_note": "",
"reference_answer_raw_annotation": "Rachel, T. Gannon, "
},
"intent_template_id": 222
}