Pith Number

pith:MSRSOI4J

pith:2026:MSRSOI4JDXJVP72R3PSD7M62RZ

not attested not anchored not stored refs resolved

Do We Really Need External Tools to Mitigate Hallucinations? SIRA: Shared-Prefix Internal Reconstruction of Attribution

Junzhe Chen, Lijie Wen, Qiang Ju, Tian Qin, Tianshu Zhang, Yuqing Shi

Masking attention to image tokens after a shared prefix in vision-language transformers reduces hallucinations by contrasting against an internal language-prior reference.

arxiv:2605.14621 v1 · 2026-05-14 · cs.CV · cs.AI · cs.CL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{MSRSOI4JDXJVP72R3PSD7M62RZ}

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

Experiments on POPE, CHAIR, and AMBER with Qwen2.5-VL and LLaVA-v1.5 show that SIRA consistently reduces hallucinations while preserving descriptive coverage and incurring lower overhead than two-pass contrastive decoding.

C2weakest assumption

That masking attention to image-token positions in later transformer layers produces a clean language-prior-dominated reference that preserves prompt interpretation and decoding history without introducing new artifacts.

C3one line summary

SIRA mitigates hallucinations in LVLMs by internally contrasting full visual access against a masked late-layer branch that retains shared context but lacks fine-grained visual evidence.

References

47 extracted · 47 resolved · 3 Pith anchors

[1] GPT-4 Technical Report 2023 · arXiv:2303.08774

[2] Mitigating object hallucinations in large vision-language models with assembly of global and local attention 2025

[3] Qwen2.5-vl technical report 2025

[4] Mask what matters: Mitigating object hallucinations in multimodal large language models with object-aligned visual contrastive decoding 2026

[5] Ict: Image-object cross-level trusted intervention for mitigating object hallucination in large vision-language models 2025

Receipt and verification

First computed	2026-05-17T23:39:04.064870Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

64a32723891dd357ff51dbe43fb3da8e43a95a41c714eb62c29d62a74e64b649

Aliases

arxiv: 2605.14621 · arxiv_version: 2605.14621v1 · doi: 10.48550/arxiv.2605.14621 · pith_short_12: MSRSOI4JDXJV · pith_short_16: MSRSOI4JDXJVP72R · pith_short_8: MSRSOI4J

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/MSRSOI4JDXJVP72R3PSD7M62RZ \
  | 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: 64a32723891dd357ff51dbe43fb3da8e43a95a41c714eb62c29d62a74e64b649

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "3cc48e668c912245f6e7b02d6a95d4a1fc4f0fac725d40509a68ba3578c63e0b",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.CL"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2026-05-14T09:37:55Z",
    "title_canon_sha256": "2afcd946870287c87c7ae834d6abf29afef5adb0660475cc7d3e2cc7e4ade6ff"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.14621",
    "kind": "arxiv",
    "version": 1
  }
}