{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:WFRDYQDRM3JRFXJQZ7GY6SYPAS","short_pith_number":"pith:WFRDYQDR","schema_version":"1.0","canonical_sha256":"b1623c407166d312dd30cfcd8f4b0f04a23ad733a43dbcedf1c22ec935fc0440","source":{"kind":"arxiv","id":"2606.17389","version":1},"attestation_state":"computed","paper":{"title":"Visuals Lie, Consistency Speaks: Disentangling Spatial Attention from Reliability in Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Ajit Saravanan, Emily Huang, Ishan Dave, Kevin Zhu, Logan Mann, Ruizhe Li, Saadullah Ismail, Shikhar Shiromani, Yi Xia","submitted_at":"2026-06-16T00:58:43Z","abstract_excerpt":"Multimodal Foundation Models are increasingly used as reasoning agents, making reliability, knowing when a model may hallucinate, critical. A common intuition, which we call the Attention-Confidence Assumption, holds that reliability follows from \"structural\" visual perception: tight attention on relevant regions should signal a trustworthy answer, while scattered attention signals confusion. We challenge this through the VLM Reliability Probe (VRP), a systematic cross-family study of reliability signals in contemporary Vision-Language Models (VLMs). We introduce structural-attention metrics, "},"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":"2606.17389","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-06-16T00:58:43Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"ee86e479ba525c8e09d4164def11ff65f11654dbd256cc4772055d700dc3e7e9","abstract_canon_sha256":"4e9440099545feec26aac7faf5e45429b2d9e95f9bd9171530ed89f0cad82b6d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:10:10.974706Z","signature_b64":"Z1xCxiHBhkSt/qiTzqyZGZ7rBIgQ50FDTG9U9c9t2I+r8Hm4nkqBsyPF3TJywo7lGH0pMEeecPu36RLNAWZ9Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b1623c407166d312dd30cfcd8f4b0f04a23ad733a43dbcedf1c22ec935fc0440","last_reissued_at":"2026-06-19T16:10:10.974362Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:10:10.974362Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Visuals Lie, Consistency Speaks: Disentangling Spatial Attention from Reliability in Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Ajit Saravanan, Emily Huang, Ishan Dave, Kevin Zhu, Logan Mann, Ruizhe Li, Saadullah Ismail, Shikhar Shiromani, Yi Xia","submitted_at":"2026-06-16T00:58:43Z","abstract_excerpt":"Multimodal Foundation Models are increasingly used as reasoning agents, making reliability, knowing when a model may hallucinate, critical. A common intuition, which we call the Attention-Confidence Assumption, holds that reliability follows from \"structural\" visual perception: tight attention on relevant regions should signal a trustworthy answer, while scattered attention signals confusion. We challenge this through the VLM Reliability Probe (VRP), a systematic cross-family study of reliability signals in contemporary Vision-Language Models (VLMs). We introduce structural-attention metrics, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.17389","kind":"arxiv","version":1},"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/2606.17389/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":"2606.17389","created_at":"2026-06-19T16:10:10.974424+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.17389v1","created_at":"2026-06-19T16:10:10.974424+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.17389","created_at":"2026-06-19T16:10:10.974424+00:00"},{"alias_kind":"pith_short_12","alias_value":"WFRDYQDRM3JR","created_at":"2026-06-19T16:10:10.974424+00:00"},{"alias_kind":"pith_short_16","alias_value":"WFRDYQDRM3JRFXJQ","created_at":"2026-06-19T16:10:10.974424+00:00"},{"alias_kind":"pith_short_8","alias_value":"WFRDYQDR","created_at":"2026-06-19T16:10:10.974424+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WFRDYQDRM3JRFXJQZ7GY6SYPAS","json":"https://pith.science/pith/WFRDYQDRM3JRFXJQZ7GY6SYPAS.json","graph_json":"https://pith.science/api/pith-number/WFRDYQDRM3JRFXJQZ7GY6SYPAS/graph.json","events_json":"https://pith.science/api/pith-number/WFRDYQDRM3JRFXJQZ7GY6SYPAS/events.json","paper":"https://pith.science/paper/WFRDYQDR"},"agent_actions":{"view_html":"https://pith.science/pith/WFRDYQDRM3JRFXJQZ7GY6SYPAS","download_json":"https://pith.science/pith/WFRDYQDRM3JRFXJQZ7GY6SYPAS.json","view_paper":"https://pith.science/paper/WFRDYQDR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.17389&json=true","fetch_graph":"https://pith.science/api/pith-number/WFRDYQDRM3JRFXJQZ7GY6SYPAS/graph.json","fetch_events":"https://pith.science/api/pith-number/WFRDYQDRM3JRFXJQZ7GY6SYPAS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WFRDYQDRM3JRFXJQZ7GY6SYPAS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WFRDYQDRM3JRFXJQZ7GY6SYPAS/action/storage_attestation","attest_author":"https://pith.science/pith/WFRDYQDRM3JRFXJQZ7GY6SYPAS/action/author_attestation","sign_citation":"https://pith.science/pith/WFRDYQDRM3JRFXJQZ7GY6SYPAS/action/citation_signature","submit_replication":"https://pith.science/pith/WFRDYQDRM3JRFXJQZ7GY6SYPAS/action/replication_record"}},"created_at":"2026-06-19T16:10:10.974424+00:00","updated_at":"2026-06-19T16:10:10.974424+00:00"}