🐦 Twitter Post Details

Viewing enriched Twitter post

@dair_ai

New research on LLM Agent Generalization. RL fine-tuning makes agents strong in familiar environments, but it struggles to transfer across unseen ones. This paper systematically studies RL generalization for LLM agents across three axes: within-environment transfer across task difficulty, cross-environment transfer to unseen settings, and sequential multi-environment training. Within an environment, RL delivers massive gains. Training on easy WebShop tasks improves hard task performance by 60+ points. Easy-to-hard curriculum learning adds another 2-3 points on top. Across environments, transfer is weak. Agents average only 3.3-3.4 point improvements on unseen environments. Training on BabyAI actually drops WebShop from 28.6 to 10.3. Sequential training is where it gets interesting. Training across five environments sequentially achieves performance comparable to joint training, with minimal forgetting. The authors claim that RL fine-tuning doesn't produce generally capable agents out of the box. But sequential training across diverse environments offers a practical path to broad competence. Paper: https://t.co/BYfVK3DPoH Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c

View on Twitter

📊 Media Metadata

{
  "media": [
    {
      "type": "photo",
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2032866111377780951/media_0.jpg",
      "filename": "media_0.jpg"
    },
    {
      "type": "photo",
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2032866111377780951/media_1.png",
      "filename": "media_1.png"
    }
  ],
  "processed_at": "2026-03-14T17:16:28.591845",
  "pipeline_version": "2.0"
}

🔧 Raw API Response

{
  "type": "tweet",
  "id": "2032866111377780951",
  "url": "https://x.com/dair_ai/status/2032866111377780951",
  "twitterUrl": "https://twitter.com/dair_ai/status/2032866111377780951",
  "text": "New research on LLM Agent Generalization.\n\nRL fine-tuning makes agents strong in familiar environments, but it struggles to transfer across unseen ones.\n\nThis paper systematically studies RL generalization for LLM agents across three axes: within-environment transfer across task difficulty, cross-environment transfer to unseen settings, and sequential multi-environment training.\n\nWithin an environment, RL delivers massive gains.\n\nTraining on easy WebShop tasks improves hard task performance by 60+ points. Easy-to-hard curriculum learning adds another 2-3 points on top.\n\nAcross environments, transfer is weak.\n\nAgents average only 3.3-3.4 point improvements on unseen environments. Training on BabyAI actually drops WebShop from 28.6 to 10.3.\n\nSequential training is where it gets interesting.\n\nTraining across five environments sequentially achieves performance comparable to joint training, with minimal forgetting.\n\nThe authors claim that RL fine-tuning doesn't produce generally capable agents out of the box.\n\nBut sequential training across diverse environments offers a practical path to broad competence.\n\nPaper: https://t.co/BYfVK3DPoH\n\nLearn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c",
  "source": "Twitter for iPhone",
  "retweetCount": 2,
  "replyCount": 1,
  "likeCount": 8,
  "quoteCount": 0,
  "viewCount": 252,
  "createdAt": "Sat Mar 14 17:07:04 +0000 2026",
  "lang": "en",
  "bookmarkCount": 7,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2032866111377780951",
  "displayTextRange": [
    0,
    279
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "dair_ai",
    "url": "https://x.com/dair_ai",
    "twitterUrl": "https://twitter.com/dair_ai",
    "id": "889050642903293953",
    "name": "DAIR.AI",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1643277398522187778/31dedbLo_normal.jpg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/889050642903293953/1773242460",
    "description": "",
    "location": "",
    "followers": 91722,
    "following": 1,
    "status": "",
    "canDm": true,
    "canMediaTag": true,
    "createdAt": "Sun Jul 23 09:12:45 +0000 2017",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 4259,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 176,
    "statusesCount": 3017,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [
      "2032459951306866714"
    ],
    "profile_bio": {
      "description": "Democratizing AI research, education, and technologies. New AI learning portal: https://t.co/LRnpZN7L4c",
      "entities": {
        "description": {
          "hashtags": [],
          "symbols": [],
          "urls": [
            {
              "display_url": "academy.dair.ai",
              "expanded_url": "https://academy.dair.ai/",
              "indices": [
                80,
                103
              ],
              "url": "https://t.co/LRnpZN7L4c"
            }
          ],
          "user_mentions": []
        },
        "url": {
          "urls": [
            {
              "display_url": "dair.ai",
              "expanded_url": "https://www.dair.ai/",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/lkqPZtMU5s"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {
    "media": [
      {
        "display_url": "pic.twitter.com/rZIiWNOfd1",
        "expanded_url": "https://twitter.com/dair_ai/status/2032866111377780951/photo/1",
        "ext_media_availability": {
          "status": "Available"
        },
        "features": {
          "large": {
            "faces": [
              {
                "h": 93,
                "w": 93,
                "x": 145,
                "y": 1312
              }
            ]
          },
          "orig": {
            "faces": [
              {
                "h": 93,
                "w": 93,
                "x": 145,
                "y": 1312
              }
            ]
          }
        },
        "id_str": "2032866107842088960",
        "indices": [
          280,
          303
        ],
        "media_key": "3_2032866107842088960",
        "media_results": {
          "id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARw2MPRZ29AACgACHDYw9SyaMNcAAA==",
          "result": {
            "__typename": "ApiMedia",
            "id": "QXBpTWVkaWE6DAABCgABHDYw9Fnb0AAKAAIcNjD1LJow1wAA",
            "media_key": "3_2032866107842088960"
          }
        },
        "media_url_https": "https://pbs.twimg.com/media/HDYw9Fnb0AAzQPW.jpg",
        "original_info": {
          "focus_rects": [
            {
              "h": 833,
              "w": 1488,
              "x": 0,
              "y": 0
            },
            {
              "h": 1488,
              "w": 1488,
              "x": 0,
              "y": 0
            },
            {
              "h": 1696,
              "w": 1488,
              "x": 0,
              "y": 0
            },
            {
              "h": 1778,
              "w": 889,
              "x": 0,
              "y": 0
            },
            {
              "h": 1778,
              "w": 1488,
              "x": 0,
              "y": 0
            }
          ],
          "height": 1778,
          "width": 1488
        },
        "sizes": {
          "large": {
            "h": 1778,
            "w": 1488
          }
        },
        "type": "photo",
        "url": "https://t.co/rZIiWNOfd1"
      }
    ]
  },
  "card": null,
  "place": {},
  "entities": {
    "hashtags": [],
    "symbols": [],
    "urls": [
      {
        "display_url": "arxiv.org/abs/2603.12011",
        "expanded_url": "https://arxiv.org/abs/2603.12011",
        "indices": [
          1126,
          1149
        ],
        "url": "https://t.co/BYfVK3DPoH"
      },
      {
        "display_url": "academy.dair.ai",
        "expanded_url": "https://academy.dair.ai/",
        "indices": [
          1202,
          1225
        ],
        "url": "https://t.co/LRnpZN7L4c"
      }
    ],
    "user_mentions": []
  },
  "quoted_tweet": null,
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "article": null
}