Pith Number

pith:CUQ74TIX

pith:2026:CUQ74TIX5EPCS7HC7HKBUAEQSY

not attested not anchored not stored refs pending

Attention Sinks in Diffusion Transformers: A Causal Analysis

Brian Summa, Fangzheng Wu

Attention sinks in diffusion transformers can be removed without degrading text-image alignment.

arxiv:2605.09313 v3 · 2026-05-10 · cs.CV

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{CUQ74TIX5EPCS7HC7HKBUAEQSY}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

removing these sinks does not degrade text-image alignment (CLIP-T) or preference proxies (ImageReward, HPS-v2) at k=1; only under stronger interventions (k≥10) does HPS-v2 exhibit a metric-dependent boundary, while CLIP-T remains robust throughout. The perceptual shifts induced by suppression are nonetheless sink-specific -- ∼6× larger than equal-budget random masking

C2weakest assumption

That the paired interventions on score and value paths causally isolate the contribution of attention sinks without introducing uncontrolled side effects on the diffusion trajectory, and that the chosen proxy metrics (CLIP-T, HPS-v2) faithfully measure semantic alignment independent of low-level perceptual style.

C3one line summary

Suppressing attention sinks in diffusion transformers does not degrade CLIP-T alignment at moderate levels but induces sink-specific perceptual shifts six times larger than equal-budget random masking.

Receipt and verification

First computed	2026-06-19T16:09:59.054972Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

1521fe4d17e91e297ce2f9d41a0090961ec2e4bd113296322c2deaa95981042b

Aliases

arxiv: 2605.09313 · arxiv_version: 2605.09313v3 · doi: 10.48550/arxiv.2605.09313 · pith_short_12: CUQ74TIX5EPC · pith_short_16: CUQ74TIX5EPCS7HC · pith_short_8: CUQ74TIX

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/CUQ74TIX5EPCS7HC7HKBUAEQSY \
  | 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: 1521fe4d17e91e297ce2f9d41a0090961ec2e4bd113296322c2deaa95981042b

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "2e678686ddf8e8f0a0a2aa85907c38763664fc2272aeaf5f5f2cbf60fb532d1f",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2026-05-10T04:14:07Z",
    "title_canon_sha256": "a76e32ef22598836394f31fe1af7964fc4df66be2d201c406ed9084348d99e7d"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.09313",
    "kind": "arxiv",
    "version": 3
  }
}