Pith Number

pith:F5X7JSFI

pith:2026:F5X7JSFI5VTT3HFJVPYXSKR33R

not attested not anchored not stored refs resolved

Revealing Interpretable Failure Modes of VLMs

Gagandeep Singh, Isha Chaudhary, Kavya Sachdeva, Sayan Ranu, Vedaant V Jain

REVELIO uncovers interpretable concept compositions that cause consistent failures in vision-language models.

arxiv:2605.12674 v1 · 2026-05-12 · cs.AI · cs.LG · cs.RO

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{F5X7JSFI5VTT3HFJVPYXSKR33R}

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

We introduce REVELIO, a framework for systematically uncovering interpretable failure modes in VLMs... uncovering previously unreported vulnerabilities in state-of-the-art VLMs.

C2weakest assumption

That the searched concept compositions correspond to genuine, consistent real-world failure modes rather than artifacts of the simulation or search heuristics.

C3one line summary

REVELIO uncovers interpretable failure modes in VLMs by searching combinatorial concept spaces with diversity-aware beam search and Gaussian-process Thompson sampling, revealing vulnerabilities in autonomous driving and indoor robotics.

References

222 extracted · 222 resolved · 3 Pith anchors

[1] Claude Sonnet 2025

[2] Large language model-assisted autonomous vehicle recovery from immobilization, 2025 2025

[3] The use of mmr, diversity-based reranking for reordering documents and producing summaries 1998

[4] Gonzalez, and Ion Stoica 2025

[5] Kevin Kai-Chun Chang, Ekin Beyazit, Alberto Sangiovanni-Vincentelli, Tichakorn Wongpirom- sarn, and Sanjit A. Seshia. Scenicrules: An autonomous driving benchmark with multi-objective specifications a 2026

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-18T03:09:50.100529Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

2f6ff4c8a8ed673d9ca9abf1792a3bdc73bd12c42f53f645a94850b3fec3b427

Aliases

arxiv: 2605.12674 · arxiv_version: 2605.12674v1 · doi: 10.48550/arxiv.2605.12674 · pith_short_12: F5X7JSFI5VTT · pith_short_16: F5X7JSFI5VTT3HFJ · pith_short_8: F5X7JSFI

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/F5X7JSFI5VTT3HFJVPYXSKR33R \
  | 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: 2f6ff4c8a8ed673d9ca9abf1792a3bdc73bd12c42f53f645a94850b3fec3b427

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "1c84029d03a85e273726e1ff54bc6df01590e0d1f25403f03d84ad349082b78c",
    "cross_cats_sorted": [
      "cs.LG",
      "cs.RO"
    ],
    "license": "http://creativecommons.org/licenses/by-nc-sa/4.0/",
    "primary_cat": "cs.AI",
    "submitted_at": "2026-05-12T19:25:17Z",
    "title_canon_sha256": "f08dd3004bd6c564804105a7d1f54d3e55105fa59590bbcacd6d209e926929a9"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.12674",
    "kind": "arxiv",
    "version": 1
  }
}