{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:KA7ICOTDHZL4TITMUSLHMEBL3D","short_pith_number":"pith:KA7ICOTD","canonical_record":{"source":{"id":"2512.21815","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2025-12-26T01:01:25Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"48f6abf74e09a01bb218402c0ee098c18be6a07cd82f4b547d10d39545aea293","abstract_canon_sha256":"87fcc8f582da8f4dfb1a9eb1534ecb945a80f771a6a8095ecc6320410f5b1e7f"},"schema_version":"1.0"},"canonical_sha256":"503e813a633e57c9a26ca49676102bd8f47736e4df6975fe74e80cc419cbfb1e","source":{"kind":"arxiv","id":"2512.21815","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2512.21815","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"arxiv_version","alias_value":"2512.21815v3","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.21815","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"pith_short_12","alias_value":"KA7ICOTDHZL4","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"pith_short_16","alias_value":"KA7ICOTDHZL4TITM","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"pith_short_8","alias_value":"KA7ICOTD","created_at":"2026-05-26T01:03:21Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:KA7ICOTDHZL4TITMUSLHMEBL3D","target":"record","payload":{"canonical_record":{"source":{"id":"2512.21815","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2025-12-26T01:01:25Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"48f6abf74e09a01bb218402c0ee098c18be6a07cd82f4b547d10d39545aea293","abstract_canon_sha256":"87fcc8f582da8f4dfb1a9eb1534ecb945a80f771a6a8095ecc6320410f5b1e7f"},"schema_version":"1.0"},"canonical_sha256":"503e813a633e57c9a26ca49676102bd8f47736e4df6975fe74e80cc419cbfb1e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T01:03:21.257956Z","signature_b64":"sUUHHuqXZzX/fRx7tA/543sZzUuioYa6M8szCh529g8uOH/PV87V8HAyctQv6mfOGrXYtE4klpORRaJslQZ+Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"503e813a633e57c9a26ca49676102bd8f47736e4df6975fe74e80cc419cbfb1e","last_reissued_at":"2026-05-26T01:03:21.256922Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T01:03:21.256922Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2512.21815","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-26T01:03:21Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3pI1G4wkZrU6biJiZSDMW8txWm6+2QKx16e26V9WL2nuJk+3O+vw0l/MQg8fN9U9NiNcuRwBUkpFQq6MCyjUDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-28T08:34:17.494953Z"},"content_sha256":"3d39be98e0d05769cf8d89f55bf1062fad4cb0b3b913ccf451e3f851d1f3ecc0","schema_version":"1.0","event_id":"sha256:3d39be98e0d05769cf8d89f55bf1062fad4cb0b3b913ccf451e3f851d1f3ecc0"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:KA7ICOTDHZL4TITMUSLHMEBL3D","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"High-Entropy Tokens as Multimodal Failure Points in Vision-Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"A small share of high-entropy tokens during generation concentrates most adversarial influence in vision-language models.","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Jing Zhang, Jinhong Ni, Mengqi He, Shu Zou, Xin Shen, Xinyu Tian, Zhaoyuan Yang","submitted_at":"2025-12-26T01:01:25Z","abstract_excerpt":"Vision-language models (VLMs) achieve remarkable performance but remain vulnerable to adversarial attacks. Entropy, as a measure of model uncertainty, is highly correlated with VLM reliability. While prior entropy-based attacks maximize uncertainty at all decoding steps, implicitly assuming that every token equally contributes to model instability, we reveal that a small fraction (around 20%) of high-entropy tokens, in the evaluated representative open-source VLMs with diverse architectures, concentrates a disproportionate share of adversarial influence during autoregressive generation. We dem"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"a small fraction (around 20%) of high-entropy tokens, in the evaluated representative open-source VLMs with diverse architectures, concentrates a disproportionate share of adversarial influence during autoregressive generation","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That high-entropy tokens can be identified reliably during generation and that concentrating perturbations on them produces comparable semantic degradation to global attacks without requiring post-hoc selection or model-specific tuning that would invalidate the transferability claim.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"High-entropy tokens act as concentrated multimodal failure points in VLMs, enabling sparse Entropy-Guided Attacks that achieve 93-95% success and 30-38% harmful rates with cross-model transfer.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A small share of high-entropy tokens during generation concentrates most adversarial influence in vision-language models.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1f8077993881cfb232972a684346efbb0dc5737bc45ee675bedbd820673b7b48"},"source":{"id":"2512.21815","kind":"arxiv","version":3},"verdict":{"id":"ec215b08-5ced-4a3d-a80c-8d4baa9b633b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T19:26:18.378293Z","strongest_claim":"a small fraction (around 20%) of high-entropy tokens, in the evaluated representative open-source VLMs with diverse architectures, concentrates a disproportionate share of adversarial influence during autoregressive generation","one_line_summary":"High-entropy tokens act as concentrated multimodal failure points in VLMs, enabling sparse Entropy-Guided Attacks that achieve 93-95% success and 30-38% harmful rates with cross-model transfer.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That high-entropy tokens can be identified reliably during generation and that concentrating perturbations on them produces comparable semantic degradation to global attacks without requiring post-hoc selection or model-specific tuning that would invalidate the transferability claim.","pith_extraction_headline":"A small share of high-entropy tokens during generation concentrates most adversarial influence in vision-language models."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2512.21815/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"b318806ae64168b412aa3aabbb7615d7f56ade2d02e8bd896fead889c1bbbaef"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"ec215b08-5ced-4a3d-a80c-8d4baa9b633b"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-26T01:03:21Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QiwzXdFGz+IKr++jguZncQQZOXbBoOivMZgiyfylHbDc593g9cktnjOVjVFMVMtE22eBVBORpCX54rGyal6OAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-28T08:34:17.495434Z"},"content_sha256":"86a42b18411cc5028a8dbf3c7ae271d0d784ba22074a53f257159a084da4718c","schema_version":"1.0","event_id":"sha256:86a42b18411cc5028a8dbf3c7ae271d0d784ba22074a53f257159a084da4718c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KA7ICOTDHZL4TITMUSLHMEBL3D/bundle.json","state_url":"https://pith.science/pith/KA7ICOTDHZL4TITMUSLHMEBL3D/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KA7ICOTDHZL4TITMUSLHMEBL3D/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-28T08:34:17Z","links":{"resolver":"https://pith.science/pith/KA7ICOTDHZL4TITMUSLHMEBL3D","bundle":"https://pith.science/pith/KA7ICOTDHZL4TITMUSLHMEBL3D/bundle.json","state":"https://pith.science/pith/KA7ICOTDHZL4TITMUSLHMEBL3D/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KA7ICOTDHZL4TITMUSLHMEBL3D/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:KA7ICOTDHZL4TITMUSLHMEBL3D","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"87fcc8f582da8f4dfb1a9eb1534ecb945a80f771a6a8095ecc6320410f5b1e7f","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2025-12-26T01:01:25Z","title_canon_sha256":"48f6abf74e09a01bb218402c0ee098c18be6a07cd82f4b547d10d39545aea293"},"schema_version":"1.0","source":{"id":"2512.21815","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2512.21815","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"arxiv_version","alias_value":"2512.21815v3","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.21815","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"pith_short_12","alias_value":"KA7ICOTDHZL4","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"pith_short_16","alias_value":"KA7ICOTDHZL4TITM","created_at":"2026-05-26T01:03:21Z"},{"alias_kind":"pith_short_8","alias_value":"KA7ICOTD","created_at":"2026-05-26T01:03:21Z"}],"graph_snapshots":[{"event_id":"sha256:86a42b18411cc5028a8dbf3c7ae271d0d784ba22074a53f257159a084da4718c","target":"graph","created_at":"2026-05-26T01:03:21Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"a small fraction (around 20%) of high-entropy tokens, in the evaluated representative open-source VLMs with diverse architectures, concentrates a disproportionate share of adversarial influence during autoregressive generation"},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That high-entropy tokens can be identified reliably during generation and that concentrating perturbations on them produces comparable semantic degradation to global attacks without requiring post-hoc selection or model-specific tuning that would invalidate the transferability claim."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"High-entropy tokens act as concentrated multimodal failure points in VLMs, enabling sparse Entropy-Guided Attacks that achieve 93-95% success and 30-38% harmful rates with cross-model transfer."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A small share of high-entropy tokens during generation concentrates most adversarial influence in vision-language models."}],"snapshot_sha256":"1f8077993881cfb232972a684346efbb0dc5737bc45ee675bedbd820673b7b48"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"b318806ae64168b412aa3aabbb7615d7f56ade2d02e8bd896fead889c1bbbaef"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2512.21815/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Vision-language models (VLMs) achieve remarkable performance but remain vulnerable to adversarial attacks. Entropy, as a measure of model uncertainty, is highly correlated with VLM reliability. While prior entropy-based attacks maximize uncertainty at all decoding steps, implicitly assuming that every token equally contributes to model instability, we reveal that a small fraction (around 20%) of high-entropy tokens, in the evaluated representative open-source VLMs with diverse architectures, concentrates a disproportionate share of adversarial influence during autoregressive generation. We dem","authors_text":"Jing Zhang, Jinhong Ni, Mengqi He, Shu Zou, Xin Shen, Xinyu Tian, Zhaoyuan Yang","cross_cats":["cs.LG"],"headline":"A small share of high-entropy tokens during generation concentrates most adversarial influence in vision-language models.","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2025-12-26T01:01:25Z","title":"High-Entropy Tokens as Multimodal Failure Points in Vision-Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2512.21815","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-16T19:26:18.378293Z","id":"ec215b08-5ced-4a3d-a80c-8d4baa9b633b","model_set":{"reader":"grok-4.3"},"one_line_summary":"High-entropy tokens act as concentrated multimodal failure points in VLMs, enabling sparse Entropy-Guided Attacks that achieve 93-95% success and 30-38% harmful rates with cross-model transfer.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A small share of high-entropy tokens during generation concentrates most adversarial influence in vision-language models.","strongest_claim":"a small fraction (around 20%) of high-entropy tokens, in the evaluated representative open-source VLMs with diverse architectures, concentrates a disproportionate share of adversarial influence during autoregressive generation","weakest_assumption":"That high-entropy tokens can be identified reliably during generation and that concentrating perturbations on them produces comparable semantic degradation to global attacks without requiring post-hoc selection or model-specific tuning that would invalidate the transferability claim."}},"verdict_id":"ec215b08-5ced-4a3d-a80c-8d4baa9b633b"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:3d39be98e0d05769cf8d89f55bf1062fad4cb0b3b913ccf451e3f851d1f3ecc0","target":"record","created_at":"2026-05-26T01:03:21Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"87fcc8f582da8f4dfb1a9eb1534ecb945a80f771a6a8095ecc6320410f5b1e7f","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2025-12-26T01:01:25Z","title_canon_sha256":"48f6abf74e09a01bb218402c0ee098c18be6a07cd82f4b547d10d39545aea293"},"schema_version":"1.0","source":{"id":"2512.21815","kind":"arxiv","version":3}},"canonical_sha256":"503e813a633e57c9a26ca49676102bd8f47736e4df6975fe74e80cc419cbfb1e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"503e813a633e57c9a26ca49676102bd8f47736e4df6975fe74e80cc419cbfb1e","first_computed_at":"2026-05-26T01:03:21.256922Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T01:03:21.256922Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"sUUHHuqXZzX/fRx7tA/543sZzUuioYa6M8szCh529g8uOH/PV87V8HAyctQv6mfOGrXYtE4klpORRaJslQZ+Dg==","signature_status":"signed_v1","signed_at":"2026-05-26T01:03:21.257956Z","signed_message":"canonical_sha256_bytes"},"source_id":"2512.21815","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3d39be98e0d05769cf8d89f55bf1062fad4cb0b3b913ccf451e3f851d1f3ecc0","sha256:86a42b18411cc5028a8dbf3c7ae271d0d784ba22074a53f257159a084da4718c"],"state_sha256":"5bfe1378217308f1bf2264ada75ebd98f3ac5af446a87b03f378164286a00606"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"apB7o8tZM1Y8TAmnYZlYEnqN9wNlLl5yCQQaU/TSFfz8ety3arftkx81boLI2JfZJ9RZ4M3bB2VVoz4KxW4gBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-28T08:34:17.497823Z","bundle_sha256":"f1fdb31069039f50862e48dbaab969a46895b894813fbee6e57e7f5f3ef5372a"}}