Viewing enriched Twitter post
Long context models are popular, but is it the final solution to long text reading? We introduce a fundamentally different method, MemWalker: 1. Build a data structure (memory tree) 2. Traverse it via LLM prompting Outperforms long context, retrieval, & recurrent baselines. (1/n) https://t.co/JrDME0ZnpB
{
"media": [
{
"id": "",
"type": "photo",
"url": null,
"media_url": "https://pbs.twimg.com/media/F8DFKoFbkAASlK4.png",
"media_url_https": null,
"display_url": null,
"expanded_url": null
}
],
"nlp": {
"sentiment": "positive",
"processed_at": "2025-08-06T12:44:45.789501"
},
"score": 1.0,
"scored_at": "2025-08-09T13:46:07.542498",
"import_source": "manual_curation_2023",
"score_components": {
"author": 0.09,
"engagement": 0.1523247582167354,
"quality": 0.18,
"source": 0.15,
"nlp": 0.1,
"recency": 0.010000000000000002
},
"source_tagged_at": "2025-08-09T13:42:53.781571",
"enriched": true,
"enriched_at": "2025-08-09T13:42:53.781572",
"links_checked": true,
"checked_at": "2025-08-10T10:32:29.505194",
"original_structure": "had_media_only"
}{
"user": {
"created_at": "2013-08-24T10:40:25.000Z",
"default_profile_image": false,
"description": "PhD student in ML/NLP @princeton_nlp & @PrincetonPLI.\nPreviously @asapp research / @cornell_tech / NTU (Taiwan).",
"fast_followers_count": 0,
"favourites_count": 3135,
"followers_count": 695,
"friends_count": 1137,
"has_custom_timelines": true,
"is_translator": false,
"listed_count": 11,
"location": "",
"media_count": 16,
"name": "Howard Chen",
"normal_followers_count": 695,
"possibly_sensitive": false,
"profile_banner_url": "https://pbs.twimg.com/profile_banners/1696122612/1662347109",
"profile_image_url_https": "https://pbs.twimg.com/profile_images/1354799599442030592/QCsWNXWD_normal.jpg",
"screen_name": "__howardchen",
"statuses_count": 114,
"translator_type": "none",
"url": "https://t.co/6KAR6m6qfv",
"verified": false,
"withheld_in_countries": [],
"id_str": "1696122612"
},
"id": "1711584916708938042",
"conversation_id": "1711584916708938042",
"full_text": "Long context models are popular, but is it the final solution to long text reading?\nWe introduce a fundamentally different method, MemWalker:\n1. Build a data structure (memory tree)\n2. Traverse it via LLM prompting\nOutperforms long context, retrieval, & recurrent baselines. (1/n) https://t.co/JrDME0ZnpB",
"reply_count": 21,
"retweet_count": 133,
"favorite_count": 846,
"hashtags": [],
"symbols": [],
"user_mentions": [],
"urls": [],
"media": [
{
"media_url": "https://pbs.twimg.com/media/F8DFKoFbkAASlK4.png",
"type": "photo"
}
],
"url": "https://twitter.com/__howardchen/status/1711584916708938042",
"created_at": "2023-10-10T03:30:18.000Z",
"#sort_index": "1711584916708938042",
"view_count": 178971,
"quote_count": 13,
"is_quote_tweet": false,
"is_retweet": false,
"is_pinned": false,
"is_truncated": false,
"startUrl": "https://twitter.com/__howardchen/status/1711584916708938042"
}