Pith Number

pith:YOAX255M

pith:2026:YOAX255MLGK57NNVGX3CZ4POSU

not attested not anchored not stored refs resolved

Senses Wide Shut: A Representation-Action Gap in Omnimodal LLMs

Fanyi Pu, Kaichen Zhang, Shuo Sun, Trung Nguyen Quang, Yiming Gao, Ziwei Liu

Omnimodal LLMs encode premise-perception mismatches in hidden states but almost never reject the conflicting claims in their outputs.

arxiv:2605.13737 v1 · 2026-05-13 · 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{YOAX255MLGK57NNVGX3CZ4POSU}

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

hidden states reliably encode premise-perception mismatches even when the same models almost never reject the false claim in their outputs

C2weakest assumption

That the IMAVB benchmark's curated misleading premises cleanly isolate grounding failures without introducing confounds from clip selection, question phrasing, or model-specific training data, and that probe-based detection of hidden-state mismatches accurately reflects functional encoding.

C3one line summary

Omnimodal LLMs encode premise-perception mismatches in hidden states yet almost never reject false textual claims, exposing a representation-action gap that is modality-asymmetric and prompt-resistant.

References

99 extracted · 99 resolved · 3 Pith anchors

[1] OpenAI GPT-5 System Card 2026 · arXiv:2601.03267

[2] A new era of intelligence with Gemini 3, 2026 2026

[3] Qwen3-omni technical report, 2025 2025

[4] Amos Azaria and Tom M. Mitchell. The internal state of an llm knows when it’s lying, 2023 2023

[5] Discovering latent knowledge in language models without supervision 2022

Receipt and verification

First computed	2026-05-18T02:44:16.505386Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

c3817d77ac5995dfb5b535f62cf1ee9511d405cbfe4962a03d36759b9c4f1026

Aliases

arxiv: 2605.13737 · arxiv_version: 2605.13737v1 · doi: 10.48550/arxiv.2605.13737 · pith_short_12: YOAX255MLGK5 · pith_short_16: YOAX255MLGK57NNV · pith_short_8: YOAX255M

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/YOAX255MLGK57NNVGX3CZ4POSU \
  | 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: c3817d77ac5995dfb5b535f62cf1ee9511d405cbfe4962a03d36759b9c4f1026

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "b0c9629294426b0a9bcec3b33ffcacda25f386e9ad045ff93db133d02dd82667",
    "cross_cats_sorted": [
      "cs.CL"
    ],
    "license": "http://creativecommons.org/licenses/by-sa/4.0/",
    "primary_cat": "cs.AI",
    "submitted_at": "2026-05-13T16:14:44Z",
    "title_canon_sha256": "7d591faafb7b2471706857ee16b2530c1ab4752fc24122f6df97aa42f4045b70"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13737",
    "kind": "arxiv",
    "version": 1
  }
}