{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreice3cqygntey3ydhh732l3yl7ef6n6ceul6hctzqnpg55fbimrow4",
"uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3medkgqdzboz2"
},
"path": "/t/seeking-arxiv-endorsement-cs-lg/173201#post_1",
"publishedAt": "2026-02-08T07:43:07.000Z",
"site": "https://discuss.huggingface.co",
"tags": [
"https://arxiv.org/auth/endorse?x=TYSM86",
"arXiv endorsement page"
],
"textContent": "Hello!\n\nI’m Dung, an independent researcher with a background in accessibility.\n\nI’m seeking your endorsement for cs.LG. My paper introduces a new metric tracking how attention heads bind multi-token terms during training. I find this signal precedes behavioral competence across Pythia checkpoints, revealing a binding-behavior decoupling at intermediate scales where internal structure and external capability diverge.\n\nThis contributes a diagnostic for detecting latent knowledge before reliable performance emerges, with implications for monitoring when internal representations may suppress capabilities.\n\nI believe the paper is a strong fit for cs.LG since it focuses on mechanistic interpretability of training dynamics and emergence in language models.\n\nYou can endorse me here: https://arxiv.org/auth/endorse?x=TYSM86\nIf the link doesn’t work, please visit arXiv endorsement page - Code: TYSM86\n\nThank you!\n\nDung.",
"title": "Seeking arXiv endorsement (cs.LG)"
}