{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:MDLVHLEXTBSVT4D75RSM7G6XMP","short_pith_number":"pith:MDLVHLEX","schema_version":"1.0","canonical_sha256":"60d753ac97986559f07fec64cf9bd763f39b27c0984157c3d0a5542c9dd598f2","source":{"kind":"arxiv","id":"2406.16320","version":3},"attestation_state":"computed","paper":{"title":"What Do VLMs NOTICE? A Mechanistic Interpretability Pipeline for Gaussian-Noise-free Text-Image Corruption and Evaluation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Carsten Eickhoff, Michal Golovanevsky, Ritambhara Singh, Vedant Palit, William Rudman","submitted_at":"2024-06-24T05:13:19Z","abstract_excerpt":"Vision-Language Models (VLMs) have gained community-spanning prominence due to their ability to integrate visual and textual inputs to perform complex tasks. Despite their success, the internal decision-making processes of these models remain opaque, posing challenges in high-stakes applications. To address this, we introduce NOTICE, the first Noise-free Text-Image Corruption and Evaluation pipeline for mechanistic interpretability in VLMs. NOTICE incorporates a Semantic Minimal Pairs (SMP) framework for image corruption and Symmetric Token Replacement (STR) for text. This approach enables sem"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2406.16320","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-24T05:13:19Z","cross_cats_sorted":[],"title_canon_sha256":"dcd6861be7de75eb14097b1126afab5a85dca6df96e142e88343607d8dd56c7d","abstract_canon_sha256":"0c9b9a7a0c0983d58d70d50ff045af21dbcb73f7a7c942795eb0012c39768ed8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:03:15.335371Z","signature_b64":"8E0PWzesNlwRw7MvMgS11op/9g7u46IOvNbVlPjreXPgF8kCbLjI5mrc6nS0tBbVC8XNR5CNT8qZgGDKiqANBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"60d753ac97986559f07fec64cf9bd763f39b27c0984157c3d0a5542c9dd598f2","last_reissued_at":"2026-07-05T11:03:15.334827Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:03:15.334827Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"What Do VLMs NOTICE? A Mechanistic Interpretability Pipeline for Gaussian-Noise-free Text-Image Corruption and Evaluation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Carsten Eickhoff, Michal Golovanevsky, Ritambhara Singh, Vedant Palit, William Rudman","submitted_at":"2024-06-24T05:13:19Z","abstract_excerpt":"Vision-Language Models (VLMs) have gained community-spanning prominence due to their ability to integrate visual and textual inputs to perform complex tasks. Despite their success, the internal decision-making processes of these models remain opaque, posing challenges in high-stakes applications. To address this, we introduce NOTICE, the first Noise-free Text-Image Corruption and Evaluation pipeline for mechanistic interpretability in VLMs. NOTICE incorporates a Semantic Minimal Pairs (SMP) framework for image corruption and Symmetric Token Replacement (STR) for text. This approach enables sem"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.16320","kind":"arxiv","version":3},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2406.16320/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":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2406.16320","created_at":"2026-07-05T11:03:15.334891+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.16320v3","created_at":"2026-07-05T11:03:15.334891+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.16320","created_at":"2026-07-05T11:03:15.334891+00:00"},{"alias_kind":"pith_short_12","alias_value":"MDLVHLEXTBSV","created_at":"2026-07-05T11:03:15.334891+00:00"},{"alias_kind":"pith_short_16","alias_value":"MDLVHLEXTBSVT4D7","created_at":"2026-07-05T11:03:15.334891+00:00"},{"alias_kind":"pith_short_8","alias_value":"MDLVHLEX","created_at":"2026-07-05T11:03:15.334891+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2602.11130","citing_title":"Meltdown: Circuits and Bifurcations in Point-Cloud-Conditioned 3D Diffusion Transformers","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07802","citing_title":"Latent Anomaly Knowledge Excavation: Unveiling Sparse Sensitive Neurons in Vision-Language Models","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04641","citing_title":"CAST: Mitigating Object Hallucination in Large Vision-Language Models via Caption-Guided Visual Attention Steering","ref_index":25,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MDLVHLEXTBSVT4D75RSM7G6XMP","json":"https://pith.science/pith/MDLVHLEXTBSVT4D75RSM7G6XMP.json","graph_json":"https://pith.science/api/pith-number/MDLVHLEXTBSVT4D75RSM7G6XMP/graph.json","events_json":"https://pith.science/api/pith-number/MDLVHLEXTBSVT4D75RSM7G6XMP/events.json","paper":"https://pith.science/paper/MDLVHLEX"},"agent_actions":{"view_html":"https://pith.science/pith/MDLVHLEXTBSVT4D75RSM7G6XMP","download_json":"https://pith.science/pith/MDLVHLEXTBSVT4D75RSM7G6XMP.json","view_paper":"https://pith.science/paper/MDLVHLEX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.16320&json=true","fetch_graph":"https://pith.science/api/pith-number/MDLVHLEXTBSVT4D75RSM7G6XMP/graph.json","fetch_events":"https://pith.science/api/pith-number/MDLVHLEXTBSVT4D75RSM7G6XMP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MDLVHLEXTBSVT4D75RSM7G6XMP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MDLVHLEXTBSVT4D75RSM7G6XMP/action/storage_attestation","attest_author":"https://pith.science/pith/MDLVHLEXTBSVT4D75RSM7G6XMP/action/author_attestation","sign_citation":"https://pith.science/pith/MDLVHLEXTBSVT4D75RSM7G6XMP/action/citation_signature","submit_replication":"https://pith.science/pith/MDLVHLEXTBSVT4D75RSM7G6XMP/action/replication_record"}},"created_at":"2026-07-05T11:03:15.334891+00:00","updated_at":"2026-07-05T11:03:15.334891+00:00"}