{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:53DB3CTIGHDFFOK6JTZ3JBTI62","short_pith_number":"pith:53DB3CTI","schema_version":"1.0","canonical_sha256":"eec61d8a6831c652b95e4cf3b48668f6bc8262f7d343a4044ee3c33c1926927c","source":{"kind":"arxiv","id":"2608.11847","version":1},"attestation_state":"computed","paper":{"title":"LookBack: Where and How to Score LVLM Responses via Visual Reference Usage","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Beomsik Cho, Dongseok Lee, Jaehyung Kim, JinHyeong Kim","submitted_at":"2026-08-12T09:40:21Z","abstract_excerpt":"Large Vision-Language Models (LVLMs) integrate visual perception with language generation, enabling responses that span image understanding and complex reasoning. However, LVLMs do not just inherit the text-level hallucinations; they also hallucinate against the image, producing fluent responses ungrounded in what they see. This makes LVLM response scoring inherently harder, and our diagnostics show that existing confidence-based metrics adopted from LLMs are insufficient for LVLMs. Specifically, removing the input image barely changes confidence-based selection, suggesting that output-space c"},"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":"2608.11847","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-08-12T09:40:21Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"a4db7a7d0c3f0bba1a92b8cf151dd0d81387a2d97925aefae333ef2bac53fbdc","abstract_canon_sha256":"a9da74b7cceaed2b6f9ee39742d96c898f64a4936aa183cc3d4b8af175d66aea"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-13T01:28:26.324115Z","signature_b64":"QayZe3k4g5AD1OhKrB+H0JlCOV30RZ4IgC7SYdSDrq9R698JBl6+l2dT5QMJrBjh650hrASkdK2E1lwwmtoUAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eec61d8a6831c652b95e4cf3b48668f6bc8262f7d343a4044ee3c33c1926927c","last_reissued_at":"2026-08-13T01:28:26.321473Z","signature_status":"signed_v1","first_computed_at":"2026-08-13T01:28:26.321473Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LookBack: Where and How to Score LVLM Responses via Visual Reference Usage","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Beomsik Cho, Dongseok Lee, Jaehyung Kim, JinHyeong Kim","submitted_at":"2026-08-12T09:40:21Z","abstract_excerpt":"Large Vision-Language Models (LVLMs) integrate visual perception with language generation, enabling responses that span image understanding and complex reasoning. However, LVLMs do not just inherit the text-level hallucinations; they also hallucinate against the image, producing fluent responses ungrounded in what they see. This makes LVLM response scoring inherently harder, and our diagnostics show that existing confidence-based metrics adopted from LLMs are insufficient for LVLMs. Specifically, removing the input image barely changes confidence-based selection, suggesting that output-space c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.11847","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/2608.11847/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":"2608.11847","created_at":"2026-08-13T01:28:26.322627+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.11847v1","created_at":"2026-08-13T01:28:26.322627+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.11847","created_at":"2026-08-13T01:28:26.322627+00:00"},{"alias_kind":"pith_short_12","alias_value":"53DB3CTIGHDF","created_at":"2026-08-13T01:28:26.322627+00:00"},{"alias_kind":"pith_short_16","alias_value":"53DB3CTIGHDFFOK6","created_at":"2026-08-13T01:28:26.322627+00:00"},{"alias_kind":"pith_short_8","alias_value":"53DB3CTI","created_at":"2026-08-13T01:28:26.322627+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/53DB3CTIGHDFFOK6JTZ3JBTI62","json":"https://pith.science/pith/53DB3CTIGHDFFOK6JTZ3JBTI62.json","graph_json":"https://pith.science/api/pith-number/53DB3CTIGHDFFOK6JTZ3JBTI62/graph.json","events_json":"https://pith.science/api/pith-number/53DB3CTIGHDFFOK6JTZ3JBTI62/events.json","paper":"https://pith.science/paper/53DB3CTI"},"agent_actions":{"view_html":"https://pith.science/pith/53DB3CTIGHDFFOK6JTZ3JBTI62","download_json":"https://pith.science/pith/53DB3CTIGHDFFOK6JTZ3JBTI62.json","view_paper":"https://pith.science/paper/53DB3CTI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.11847&json=true","fetch_graph":"https://pith.science/api/pith-number/53DB3CTIGHDFFOK6JTZ3JBTI62/graph.json","fetch_events":"https://pith.science/api/pith-number/53DB3CTIGHDFFOK6JTZ3JBTI62/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/53DB3CTIGHDFFOK6JTZ3JBTI62/action/timestamp_anchor","attest_storage":"https://pith.science/pith/53DB3CTIGHDFFOK6JTZ3JBTI62/action/storage_attestation","attest_author":"https://pith.science/pith/53DB3CTIGHDFFOK6JTZ3JBTI62/action/author_attestation","sign_citation":"https://pith.science/pith/53DB3CTIGHDFFOK6JTZ3JBTI62/action/citation_signature","submit_replication":"https://pith.science/pith/53DB3CTIGHDFFOK6JTZ3JBTI62/action/replication_record"}},"created_at":"2026-08-13T01:28:26.322627+00:00","updated_at":"2026-08-13T01:28:26.322627+00:00"}