Viewing enriched Twitter post
Evaluating LLM outputs can be time-consuming and challenging even for experts. That's where LLM-as-a-Judge comes in. We believe all AI devs should get familiar with this technique. Here is why: LLM-as-a-Judge, automates the assessment of LLM outputs by using a specialized LLM… https://t.co/3FML9UMhjq
{
"score": 0.83,
"scored_at": "2025-08-09T13:47:19.523179",
"import_source": "unknown_source",
"links_checked": true,
"checked_at": "2025-08-10T10:33:19.348538",
"original_structure": "had_data_only",
"media": [
{
"id": "1903098087058575361",
"type": "photo",
"url": "https://t.co/3FML9UMhjq",
"media_url": null,
"media_url_https": "https://pbs.twimg.com/media/GmkpoIvaEAETqG9.png",
"display_url": "pic.x.com/3FML9UMhjq",
"expanded_url": "https://x.com/dair_ai/status/1903098701440061592/photo/1"
}
]
}{
"tweet": {
"bookmark_count": 186,
"bookmarked": false,
"created_at": "Fri Mar 21 14:57:42 +0000 2025",
"conversation_id_str": "1903098701440061592",
"display_text_range": [
0,
281
],
"entities": {
"hashtags": [],
"media": [
{
"display_url": "pic.x.com/3FML9UMhjq",
"expanded_url": "https://x.com/dair_ai/status/1903098701440061592/photo/1",
"id_str": "1903098087058575361",
"indices": [
282,
305
],
"media_key": "3_1903098087058575361",
"media_url_https": "https://pbs.twimg.com/media/GmkpoIvaEAETqG9.png",
"type": "photo",
"url": "https://t.co/3FML9UMhjq",
"ext_media_availability": {
"status": "Available"
},
"features": {
"large": {
"faces": []
},
"medium": {
"faces": []
},
"small": {
"faces": []
},
"orig": {
"faces": []
}
},
"sizes": {
"large": {
"h": 1608,
"w": 1628,
"resize": "fit"
},
"medium": {
"h": 1185,
"w": 1200,
"resize": "fit"
},
"small": {
"h": 672,
"w": 680,
"resize": "fit"
},
"thumb": {
"h": 150,
"w": 150,
"resize": "crop"
}
},
"original_info": {
"height": 1608,
"width": 1628,
"focus_rects": [
{
"x": 0,
"y": 398,
"w": 1628,
"h": 912
},
{
"x": 0,
"y": 0,
"w": 1608,
"h": 1608
},
{
"x": 0,
"y": 0,
"w": 1411,
"h": 1608
},
{
"x": 0,
"y": 0,
"w": 804,
"h": 1608
},
{
"x": 0,
"y": 0,
"w": 1628,
"h": 1608
}
]
},
"allow_download_status": {
"allow_download": true
},
"media_results": {
"result": {
"media_key": "3_1903098087058575361"
}
}
}
],
"symbols": [],
"timestamps": [],
"urls": [],
"user_mentions": []
},
"extended_entities": {
"media": [
{
"display_url": "pic.x.com/3FML9UMhjq",
"expanded_url": "https://x.com/dair_ai/status/1903098701440061592/photo/1",
"id_str": "1903098087058575361",
"indices": [
282,
305
],
"media_key": "3_1903098087058575361",
"media_url_https": "https://pbs.twimg.com/media/GmkpoIvaEAETqG9.png",
"type": "photo",
"url": "https://t.co/3FML9UMhjq",
"ext_media_availability": {
"status": "Available"
},
"features": {
"large": {
"faces": []
},
"medium": {
"faces": []
},
"small": {
"faces": []
},
"orig": {
"faces": []
}
},
"sizes": {
"large": {
"h": 1608,
"w": 1628,
"resize": "fit"
},
"medium": {
"h": 1185,
"w": 1200,
"resize": "fit"
},
"small": {
"h": 672,
"w": 680,
"resize": "fit"
},
"thumb": {
"h": 150,
"w": 150,
"resize": "crop"
}
},
"original_info": {
"height": 1608,
"width": 1628,
"focus_rects": [
{
"x": 0,
"y": 398,
"w": 1628,
"h": 912
},
{
"x": 0,
"y": 0,
"w": 1608,
"h": 1608
},
{
"x": 0,
"y": 0,
"w": 1411,
"h": 1608
},
{
"x": 0,
"y": 0,
"w": 804,
"h": 1608
},
{
"x": 0,
"y": 0,
"w": 1628,
"h": 1608
}
]
},
"allow_download_status": {
"allow_download": true
},
"media_results": {
"result": {
"media_key": "3_1903098087058575361"
}
}
}
]
},
"favorite_count": 160,
"favorited": false,
"full_text": "Evaluating LLM outputs can be time-consuming and challenging even for experts.\n\nThat's where LLM-as-a-Judge comes in.\n\nWe believe all AI devs should get familiar with this technique.\n\nHere is why:\n\nLLM-as-a-Judge, automates the assessment of LLM outputs by using a specialized LLM… https://t.co/3FML9UMhjq",
"is_quote_status": false,
"lang": "en",
"possibly_sensitive": false,
"possibly_sensitive_editable": true,
"quote_count": 1,
"reply_count": 1,
"retweet_count": 19,
"retweeted": false,
"user_id_str": "889050642903293953",
"id_str": "1903098701440061592",
"note_tweet": {
"is_expandable": true,
"note_tweet_results": {
"result": {
"id": "Tm90ZVR3ZWV0OjE5MDMwOTg3MDEyNzY0MDE2NjQ=",
"text": "Evaluating LLM outputs can be time-consuming and challenging even for experts.\n\nThat's where LLM-as-a-Judge comes in.\n\nWe believe all AI devs should get familiar with this technique.\n\nHere is why:\n\nLLM-as-a-Judge, automates the assessment of LLM outputs by using a specialized LLM as a “Judge”.\n\nThis enables rapid development of LLM applications and AI agents as response assessment can now be automated with LLMs. \n\nReasoning models like o3-mini and o1 are ideal to implement a LLM-as-a-Judge system due to their ability to reason and understand more complex data and follow advanced requirements. \n\nHere’s how you can set up an LLM-as-a-Judge system to automatically score outputs on criteria like accuracy, relevance, and tone:\n\n1. Define Evaluation Criteria: Create examples of “good” and “bad” responses in your evaluation prompt.\n2. Automate Scoring: Integrate the judge LLM into your pipeline so every output is instantly scored with brief feedback.\n3. Real-Time Monitoring: Use these scores to track quality over time, optimize prompts, and ensure responses stay faithful to their context.\n\nLLM-as-a-Judge is increasingly being used in meta-prompting workflows. Using a powerful reasoning model like o1-mini or o3-mini to optimize prompts for chat models is a subject we will explore in another post later this week.\n\nOur Prompt Engineering for Developers course covers both subjects, with code examples devs can use in their own projects! Check out the link below ⬇️",
"entity_set": {
"hashtags": [],
"symbols": [],
"urls": [],
"user_mentions": []
},
"richtext": {
"richtext_tags": []
},
"media": {
"inline_media": []
}
}
}
}
},
"user": {
"__typename": "User",
"id": "VXNlcjo4ODkwNTA2NDI5MDMyOTM5NTM=",
"rest_id": "889050642903293953",
"affiliates_highlighted_label": {},
"has_graduated_access": true,
"is_blue_verified": true,
"profile_image_shape": "Circle",
"legacy": {
"can_dm": true,
"can_media_tag": true,
"created_at": "Sun Jul 23 09:12:45 +0000 2017",
"default_profile": false,
"default_profile_image": false,
"description": "Democratizing AI research, education, and technologies. Learn how to build with AI in our new AI Academy: https://t.co/zQXQt0Pem8",
"entities": {
"description": {
"urls": [
{
"display_url": "dair-ai.thinkific.com",
"expanded_url": "https://dair-ai.thinkific.com/",
"url": "https://t.co/zQXQt0Pem8",
"indices": [
106,
129
]
}
]
},
"url": {
"urls": [
{
"display_url": "github.com/dair-ai",
"expanded_url": "https://github.com/dair-ai",
"url": "https://t.co/W6313m1okg",
"indices": [
0,
23
]
}
]
}
},
"fast_followers_count": 0,
"favourites_count": 3160,
"followers_count": 73952,
"friends_count": 1,
"has_custom_timelines": true,
"is_translator": false,
"listed_count": 1290,
"location": "",
"media_count": 64,
"name": "DAIR.AI",
"normal_followers_count": 73952,
"pinned_tweet_ids_str": [
"1936825775304634798"
],
"possibly_sensitive": false,
"profile_banner_url": "https://pbs.twimg.com/profile_banners/889050642903293953/1742055232",
"profile_image_url_https": "https://pbs.twimg.com/profile_images/1643277398522187778/31dedbLo_normal.jpg",
"profile_interstitial_type": "",
"screen_name": "dair_ai",
"statuses_count": 2059,
"translator_type": "none",
"url": "https://t.co/W6313m1okg",
"verified": false,
"want_retweets": false,
"withheld_in_countries": []
},
"professional": {
"rest_id": "1548740760790507520",
"professional_type": "Business",
"category": []
},
"tipjar_settings": {
"is_enabled": false
}
},
"views": "26634"
}