{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreidzyc2yxebl3hgez4lkyp2ixeqbo4glr7kqfagbr6pya3vlmbpg64",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mmyii3dp5no2"
  },
  "path": "/t/attention-is-all-we-had-but-not-what-we-needed-language-generation-without-attention-via-iterative-energy-based-state-refinement/176285#post_11",
  "publishedAt": "2026-05-29T10:05:37.000Z",
  "site": "https://discuss.huggingface.co",
  "tags": [
    "Zenodo",
    "Attention Is All We Had — But Not What We Needed: Convergent State Machine..."
  ],
  "textContent": "New Version\n\nZenodo\n\n### Attention Is All We Had — But Not What We Needed: Convergent State Machine...\n\nWe introduce the Convergent State Machine (CSM), a novel architecture for language generation that replaces attention entirely with energy-based iterative state refinement. Three models trained (66M, 150M, 331M), all with zero attention layers. CSM...",
  "title": "Attention Is All We Had — But Not What We Needed: Language generation without attention via iterative energy-based state refinement"
}