pith:MWMWF35W
RAT+: Train Dense, Infer Sparse -- Recurrence Augmented Attention for Dilated Inference
RAT+ lets one densely pretrained model switch to dilated sparse attention at inference with only short adaptation.
arxiv:2602.18196 v4 · 2026-02-20 · cs.LG
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{MWMWF35W6EB7PUGEM7D64EVO2J}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
A single RAT+ model is pretrained densely once and can then be flexibly switched at inference time to dilated attention (optionally with local windows) or hybrid layer/head compositions, requiring only a short 1B-token resolution adaptation rather than retraining separate sparse models.
That adding full-sequence recurrence and active recurrence learning during dense pretraining creates representations that transfer to dilated sparse patterns with only short adaptation and limited accuracy loss.
RAT+ pretrains a single dense recurrent-augmented attention model that supports flexible dilated sparse inference after short adaptation, matching dense accuracy at moderate dilation and losing only 1-3 points at high dilation.
Receipt and verification
| First computed | 2026-05-21T01:05:16.476570Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
659962efb6f103f7d0c467c7ee12aed24291f1daa1f6c17cfa39ee8081e6df52
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/MWMWF35W6EB7PUGEM7D64EVO2J \
| jq -c '.canonical_record' \
| python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 659962efb6f103f7d0c467c7ee12aed24291f1daa1f6c17cfa39ee8081e6df52
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "0e8deeea39032e68e54efa8dcbc3038a25f3c084329d43f21b6c526445c12029",
"cross_cats_sorted": [],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"primary_cat": "cs.LG",
"submitted_at": "2026-02-20T13:09:49Z",
"title_canon_sha256": "62e201d3d163f9d7ceb6c2b5c7300780dfead246ea2ecf0c91a821de8f2a5ffb"
},
"schema_version": "1.0",
"source": {
"id": "2602.18196",
"kind": "arxiv",
"version": 4
}
}