{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:H7TB5IEFHLF6SMKSEFTXBYZXZI","short_pith_number":"pith:H7TB5IEF","schema_version":"1.0","canonical_sha256":"3fe61ea0853acbe93152216770e337ca1f243d93e806709592e7c26dd82a73c1","source":{"kind":"arxiv","id":"2602.15892","version":2},"attestation_state":"computed","paper":{"title":"Egocentric Bias in Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bingyang Wang, Dezhi Luo, Emmy Liu, Hokin Deng, Maijunxian Wang, Qingying Gao, Ran Ji, Tianwei Zhao, Yijiang Li","submitted_at":"2026-02-10T03:51:00Z","abstract_excerpt":"Visual perspective taking--inferring how the world appears from another's viewpoint--is foundational to social cognition. We introduce FlipSet, a diagnostic benchmark for Level-2 visual perspective taking (L2 VPT) in vision-language models. The task requires simulating 180-degree rotations of 2D character strings from another agent's perspective, isolating spatial transformation from 3D scene complexity. Evaluating 103 VLMs reveals systematic egocentric bias: the vast majority perform below chance, with roughly three-quarters of errors reproducing the camera viewpoint. Control experiments expo"},"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":"2602.15892","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-02-10T03:51:00Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"6f2f841e9a26bd80a5f81b25b3b4397581940fbe4e86e5540b9a7ac5e1a41d01","abstract_canon_sha256":"d68b41457fd1d8c341074bb71f2e742308e0c61c8e2997ee031914a75473afb3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-15T01:21:50.622463Z","signature_b64":"qEYYZHqYPU4HZNLn0wBy+Z3UWbYiJ5VVIM6+vh83/0FT8X4CgWyT8fcenygigK9iZYD8EgL2KEH21IsZ74PyAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3fe61ea0853acbe93152216770e337ca1f243d93e806709592e7c26dd82a73c1","last_reissued_at":"2026-07-15T01:21:50.621430Z","signature_status":"signed_v1","first_computed_at":"2026-07-15T01:21:50.621430Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Egocentric Bias in Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bingyang Wang, Dezhi Luo, Emmy Liu, Hokin Deng, Maijunxian Wang, Qingying Gao, Ran Ji, Tianwei Zhao, Yijiang Li","submitted_at":"2026-02-10T03:51:00Z","abstract_excerpt":"Visual perspective taking--inferring how the world appears from another's viewpoint--is foundational to social cognition. We introduce FlipSet, a diagnostic benchmark for Level-2 visual perspective taking (L2 VPT) in vision-language models. The task requires simulating 180-degree rotations of 2D character strings from another agent's perspective, isolating spatial transformation from 3D scene complexity. Evaluating 103 VLMs reveals systematic egocentric bias: the vast majority perform below chance, with roughly three-quarters of errors reproducing the camera viewpoint. Control experiments expo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.15892","kind":"arxiv","version":2},"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/2602.15892/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":"2602.15892","created_at":"2026-07-15T01:21:50.621894+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.15892v2","created_at":"2026-07-15T01:21:50.621894+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.15892","created_at":"2026-07-15T01:21:50.621894+00:00"},{"alias_kind":"pith_short_12","alias_value":"H7TB5IEFHLF6","created_at":"2026-07-15T01:21:50.621894+00:00"},{"alias_kind":"pith_short_16","alias_value":"H7TB5IEFHLF6SMKS","created_at":"2026-07-15T01:21:50.621894+00:00"},{"alias_kind":"pith_short_8","alias_value":"H7TB5IEF","created_at":"2026-07-15T01:21:50.621894+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":5,"sample":[{"citing_arxiv_id":"2605.23176","citing_title":"DRIVESPATIAL: A Benchmark for Spatiotemporal Intelligence in VLMs for Autonomous Driving","ref_index":70,"is_internal_anchor":true},{"citing_arxiv_id":"2605.23176","citing_title":"DRIVESPATIAL: A Benchmark for Spatiotemporal Intelligence in VLMs for Autonomous Driving","ref_index":70,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12413","citing_title":"Beyond Localization: A Comprehensive Diagnosis of Perspective-Conditioned Spatial Reasoning in MLLMs from Omnidirectional Images","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12413","citing_title":"Beyond Localization: A Comprehensive Diagnosis of Perspective-Conditioned Spatial Reasoning in MLLMs from Omnidirectional Images","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12413","citing_title":"Beyond Localization: A Comprehensive Diagnosis of Perspective-Conditioned Spatial Reasoning in MLLMs from Omnidirectional Images","ref_index":48,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/H7TB5IEFHLF6SMKSEFTXBYZXZI","json":"https://pith.science/pith/H7TB5IEFHLF6SMKSEFTXBYZXZI.json","graph_json":"https://pith.science/api/pith-number/H7TB5IEFHLF6SMKSEFTXBYZXZI/graph.json","events_json":"https://pith.science/api/pith-number/H7TB5IEFHLF6SMKSEFTXBYZXZI/events.json","paper":"https://pith.science/paper/H7TB5IEF"},"agent_actions":{"view_html":"https://pith.science/pith/H7TB5IEFHLF6SMKSEFTXBYZXZI","download_json":"https://pith.science/pith/H7TB5IEFHLF6SMKSEFTXBYZXZI.json","view_paper":"https://pith.science/paper/H7TB5IEF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.15892&json=true","fetch_graph":"https://pith.science/api/pith-number/H7TB5IEFHLF6SMKSEFTXBYZXZI/graph.json","fetch_events":"https://pith.science/api/pith-number/H7TB5IEFHLF6SMKSEFTXBYZXZI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/H7TB5IEFHLF6SMKSEFTXBYZXZI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/H7TB5IEFHLF6SMKSEFTXBYZXZI/action/storage_attestation","attest_author":"https://pith.science/pith/H7TB5IEFHLF6SMKSEFTXBYZXZI/action/author_attestation","sign_citation":"https://pith.science/pith/H7TB5IEFHLF6SMKSEFTXBYZXZI/action/citation_signature","submit_replication":"https://pith.science/pith/H7TB5IEFHLF6SMKSEFTXBYZXZI/action/replication_record"}},"created_at":"2026-07-15T01:21:50.621894+00:00","updated_at":"2026-07-15T01:21:50.621894+00:00"}