e433f35c1f
- Add `fixed_ip` field to all 369 JSON files in examples directory - Set to `true` for 8 files listed in google_chrome.json multi_apps - Set to `false` for remaining 361 files - Add `possibility_of_env_change` field to 363 JSON files missing this field - Set to "low" for newly added fields - Preserve existing values (4 medium, 2 high) for 6 files that already had this field This ensures consistent configuration schema across all evaluation examples while maintaining backward compatibility with existing settings.
73 lines
2.3 KiB
JSON
73 lines
2.3 KiB
JSON
{
|
|
"id": "5bc63fb9-276a-4439-a7c1-9dc76401737f",
|
|
"snapshot": "multiapps",
|
|
"instruction": "I have a JSON-formatted data file opened now that stores the responses of several large language models on a specific dataset. Now, I want to filter out all the responses from Gemini and specifically look at the sentences in the responses that contain \"Iliad\". Please copy all of Gemini's responses(values) from the JSON file, paste them into a Word document named \"gemini_results.docx\" under Desktop. Each response should be a new paragraph and sepreated by a new line. Highlight the all the \"Iliad\" word.",
|
|
"source": "",
|
|
"config": [
|
|
{
|
|
"type": "download",
|
|
"parameters": {
|
|
"files": [
|
|
{
|
|
"url": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/multi_apps/5bc63fb9-276a-4439-a7c1-9dc76401737f/LLM%20results.json",
|
|
"path": "/home/user/Desktop/llm_answers.json"
|
|
}
|
|
]
|
|
}
|
|
},
|
|
{
|
|
"type": "open",
|
|
"parameters": {
|
|
"path": "/home/user/Desktop/llm_answers.json"
|
|
}
|
|
}
|
|
],
|
|
"trajectory": "trajectories/",
|
|
"related_apps": [
|
|
"libreoffice_writer"
|
|
],
|
|
"evaluator": {
|
|
"postconfig": [
|
|
{
|
|
"type": "activate_window",
|
|
"parameters": {
|
|
"window_name": "gemini_results.docx - LibreOffice Writer",
|
|
"strict": true
|
|
}
|
|
},
|
|
{
|
|
"type": "sleep",
|
|
"parameters": {
|
|
"seconds": 0.5
|
|
}
|
|
},
|
|
{
|
|
"type": "execute",
|
|
"parameters": {
|
|
"command": [
|
|
"python",
|
|
"-c",
|
|
"import pyautogui; import time; pyautogui.hotkey('ctrl', 's'); time.sleep(0.5); "
|
|
]
|
|
}
|
|
}
|
|
],
|
|
"func": "compare_docx_files_and_ignore_new_lines",
|
|
"expected": {
|
|
"type": "cloud_file",
|
|
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/multi_apps/5bc63fb9-276a-4439-a7c1-9dc76401737f/gemini_results_Gold.docx",
|
|
"dest": "gemini_results_Gold.docx"
|
|
},
|
|
"result": {
|
|
"type": "vm_file",
|
|
"path": "/home/user/Desktop/gemini_results.docx",
|
|
"dest": "gemini_results.docx"
|
|
},
|
|
"options": {
|
|
"ignore_blanks": false
|
|
}
|
|
},
|
|
"proxy": false,
|
|
"fixed_ip": false,
|
|
"possibility_of_env_change": "low"
|
|
} |