🐦 Twitter Post Details

Viewing enriched Twitter post

@giffmana

An aux task for LLM to predict the order of next chunk of tokens. Haha this is so 2018! In computer vision this kind of tasks was all the rage for "self-supervised" learning! This specific one was called "jigsaw". There was also "relative patch location" and "rotation prediction". All follow the same pattern: you artificially mess with the image, and the model needs to find out what you did. Actually, this was my first paper with @__kolesnikov__ and @XiaohuaZhai , where we basically said: "wait a minute, all of these aux tasks are kind of a big mess" (while also setting new sota using them, so we are allowed to say that lol) It's interesting because i see more and more previous vision things becoming relevant to language modeling lately, i love it!

Media 1

📊 Media Metadata

{
  "media": [
    {
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1960763690891829402/media_0.jpg?",
      "media_url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1960763690891829402/media_0.jpg?",
      "type": "photo",
      "filename": "media_0.jpg"
    }
  ],
  "processed_at": "2025-08-28T13:34:20.518616",
  "pipeline_version": "2.0"
}

🔧 Raw API Response

{
  "type": "tweet",
  "id": "1960763690891829402",
  "url": "https://x.com/giffmana/status/1960763690891829402",
  "twitterUrl": "https://twitter.com/giffmana/status/1960763690891829402",
  "text": "An aux task for LLM to predict the order of next chunk of tokens. Haha this is so 2018!\n\nIn computer vision this kind of tasks was all the rage for \"self-supervised\" learning! This specific one was called \"jigsaw\". There was also \"relative patch location\" and \"rotation prediction\". All follow the same pattern: you artificially mess with the image, and the model needs to find out what you did.\n\nActually, this was my first paper with @__kolesnikov__ and @XiaohuaZhai , where we basically said: \"wait a minute, all of these aux tasks are kind of a big mess\" (while also setting new sota using them, so we are allowed to say that lol)\n\nIt's interesting because i see more and more previous vision things becoming relevant to language modeling lately, i love it!",
  "source": "Twitter for iPhone",
  "retweetCount": 17,
  "replyCount": 15,
  "likeCount": 311,
  "quoteCount": 3,
  "viewCount": 57208,
  "createdAt": "Wed Aug 27 17:57:47 +0000 2025",
  "lang": "en",
  "bookmarkCount": 203,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "1960763690891829402",
  "displayTextRange": [
    0,
    270
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "giffmana",
    "url": "https://x.com/giffmana",
    "twitterUrl": "https://twitter.com/giffmana",
    "id": "2236047510",
    "name": "Lucas Beyer (bl16)",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/378800000845687873/37bba4f807fe3a2c644a252f8191338d_normal.jpeg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/2236047510/1548960441",
    "description": "",
    "location": "Zürich, Suisse",
    "followers": 107113,
    "following": 516,
    "status": "",
    "canDm": false,
    "canMediaTag": false,
    "createdAt": "Sun Dec 08 13:31:09 +0000 2013",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 50550,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 1985,
    "statusesCount": 21922,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [
      "1570152923233144832"
    ],
    "profile_bio": {
      "description": "Researcher (now: Meta. ex:  OpenAI, DeepMind, Brain, RWTH Aachen), Gamer, Hacker, Belgian. Anon feedback: https://t.co/xe2XUqkKit\n✗DMs → email",
      "entities": {
        "description": {
          "urls": [
            {
              "display_url": "admonymous.co/giffmana",
              "expanded_url": "https://www.admonymous.co/giffmana",
              "indices": [
                106,
                129
              ],
              "url": "https://t.co/xe2XUqkKit"
            }
          ]
        },
        "url": {
          "urls": [
            {
              "display_url": "lucasb.eyer.be",
              "expanded_url": "http://lucasb.eyer.be",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/RsCh9TJjKC"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {
    "media": [
      {
        "display_url": "pic.twitter.com/k7SJShfEWg",
        "expanded_url": "https://twitter.com/giffmana/status/1960763690891829402/photo/1",
        "ext_media_availability": {
          "status": "Available"
        },
        "features": {
          "large": {},
          "orig": {}
        },
        "id_str": "1960763683899932672",
        "indices": [
          271,
          294
        ],
        "media_key": "3_1960763683899932672",
        "media_results": {
          "id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARs2CC6TV4AACgACGzYIMDQXYJoAAA==",
          "result": {
            "__typename": "ApiMedia",
            "id": "QXBpTWVkaWE6DAABCgABGzYILpNXgAAKAAIbNggwNBdgmgAA",
            "media_key": "3_1960763683899932672"
          }
        },
        "media_url_https": "https://pbs.twimg.com/media/GzYILpNXgAAsZMO.jpg",
        "original_info": {
          "focus_rects": [
            {
              "h": 675,
              "w": 1206,
              "x": 0,
              "y": 0
            },
            {
              "h": 774,
              "w": 774,
              "x": 246,
              "y": 0
            },
            {
              "h": 774,
              "w": 679,
              "x": 294,
              "y": 0
            },
            {
              "h": 774,
              "w": 387,
              "x": 440,
              "y": 0
            },
            {
              "h": 774,
              "w": 1206,
              "x": 0,
              "y": 0
            }
          ],
          "height": 774,
          "width": 1206
        },
        "sizes": {
          "large": {
            "h": 774,
            "w": 1206
          }
        },
        "type": "photo",
        "url": "https://t.co/k7SJShfEWg"
      }
    ]
  },
  "card": null,
  "place": {},
  "entities": {
    "user_mentions": [
      {
        "id_str": "1091042849561473031",
        "indices": [
          436,
          451
        ],
        "name": "Alexander Kolesnikov",
        "screen_name": "__kolesnikov__"
      },
      {
        "id_str": "1250971237",
        "indices": [
          456,
          468
        ],
        "name": "Xiaohua Zhai",
        "screen_name": "XiaohuaZhai"
      }
    ]
  },
  "quoted_tweet": {
    "type": "tweet",
    "id": "1960642776577794277",
    "url": "",
    "twitterUrl": "",
    "text": "",
    "source": "Twitter for iPhone",
    "retweetCount": 0,
    "replyCount": 0,
    "likeCount": 0,
    "quoteCount": 0,
    "viewCount": 0,
    "createdAt": "",
    "lang": "",
    "bookmarkCount": 0,
    "isReply": false,
    "inReplyToId": null,
    "conversationId": "",
    "displayTextRange": [],
    "inReplyToUserId": null,
    "inReplyToUsername": null,
    "author": {},
    "extendedEntities": {},
    "card": null,
    "place": {},
    "entities": {},
    "quoted_tweet": null,
    "retweeted_tweet": null,
    "isLimitedReply": false,
    "article": null
  },
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "article": null
}