{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CX63BD4Q4FBYXK5RSX5Y7A2ZQE","short_pith_number":"pith:CX63BD4Q","schema_version":"1.0","canonical_sha256":"15fdb08f90e1438babb195fb8f83598109febc984e768323afc4084dbbf20eaa","source":{"kind":"arxiv","id":"2508.06525","version":1},"attestation_state":"computed","paper":{"title":"Large Language Models Facilitate Vision Reflection in Image Classification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guoyuan An, Jaeyoon Kim, Sungeui Yoon","submitted_at":"2025-08-02T03:34:42Z","abstract_excerpt":"This paper presents several novel findings on the explainability of vision reflection in large multimodal models (LMMs). First, we show that prompting an LMM to verify the prediction of a specialized vision model can improve recognition accuracy, even on benchmarks like ImageNet, despite prior evidence that LMMs typically underperform dedicated vision encoders. Second, we analyze the internal behavior of vision reflection and find that the vision-language connector maps visual features into explicit textual concepts, allowing the language model to reason about prediction plausibility using com"},"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":"2508.06525","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-08-02T03:34:42Z","cross_cats_sorted":[],"title_canon_sha256":"ad44a930be712632cbe671e55d73a3903a19b7e1fc12a3c3be2163730c329e45","abstract_canon_sha256":"914aed1531c73e61c42008b005364a0e84865f03ed5e32bbcdddd5c36ac6a356"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:51:01.174011Z","signature_b64":"jsNAfrLhqs+yK1zePLy9Y0vI92yhLPkFjRobrfaC/n8oWwYsH8jUnrFpHxtP1/MIPiHB9psqJVHZzSio9QFIBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"15fdb08f90e1438babb195fb8f83598109febc984e768323afc4084dbbf20eaa","last_reissued_at":"2026-07-05T11:51:01.173476Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:51:01.173476Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Large Language Models Facilitate Vision Reflection in Image Classification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guoyuan An, Jaeyoon Kim, Sungeui Yoon","submitted_at":"2025-08-02T03:34:42Z","abstract_excerpt":"This paper presents several novel findings on the explainability of vision reflection in large multimodal models (LMMs). First, we show that prompting an LMM to verify the prediction of a specialized vision model can improve recognition accuracy, even on benchmarks like ImageNet, despite prior evidence that LMMs typically underperform dedicated vision encoders. Second, we analyze the internal behavior of vision reflection and find that the vision-language connector maps visual features into explicit textual concepts, allowing the language model to reason about prediction plausibility using com"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.06525","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/2508.06525/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":"2508.06525","created_at":"2026-07-05T11:51:01.173534+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.06525v1","created_at":"2026-07-05T11:51:01.173534+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.06525","created_at":"2026-07-05T11:51:01.173534+00:00"},{"alias_kind":"pith_short_12","alias_value":"CX63BD4Q4FBY","created_at":"2026-07-05T11:51:01.173534+00:00"},{"alias_kind":"pith_short_16","alias_value":"CX63BD4Q4FBYXK5R","created_at":"2026-07-05T11:51:01.173534+00:00"},{"alias_kind":"pith_short_8","alias_value":"CX63BD4Q","created_at":"2026-07-05T11:51:01.173534+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/CX63BD4Q4FBYXK5RSX5Y7A2ZQE","json":"https://pith.science/pith/CX63BD4Q4FBYXK5RSX5Y7A2ZQE.json","graph_json":"https://pith.science/api/pith-number/CX63BD4Q4FBYXK5RSX5Y7A2ZQE/graph.json","events_json":"https://pith.science/api/pith-number/CX63BD4Q4FBYXK5RSX5Y7A2ZQE/events.json","paper":"https://pith.science/paper/CX63BD4Q"},"agent_actions":{"view_html":"https://pith.science/pith/CX63BD4Q4FBYXK5RSX5Y7A2ZQE","download_json":"https://pith.science/pith/CX63BD4Q4FBYXK5RSX5Y7A2ZQE.json","view_paper":"https://pith.science/paper/CX63BD4Q","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.06525&json=true","fetch_graph":"https://pith.science/api/pith-number/CX63BD4Q4FBYXK5RSX5Y7A2ZQE/graph.json","fetch_events":"https://pith.science/api/pith-number/CX63BD4Q4FBYXK5RSX5Y7A2ZQE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CX63BD4Q4FBYXK5RSX5Y7A2ZQE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CX63BD4Q4FBYXK5RSX5Y7A2ZQE/action/storage_attestation","attest_author":"https://pith.science/pith/CX63BD4Q4FBYXK5RSX5Y7A2ZQE/action/author_attestation","sign_citation":"https://pith.science/pith/CX63BD4Q4FBYXK5RSX5Y7A2ZQE/action/citation_signature","submit_replication":"https://pith.science/pith/CX63BD4Q4FBYXK5RSX5Y7A2ZQE/action/replication_record"}},"created_at":"2026-07-05T11:51:01.173534+00:00","updated_at":"2026-07-05T11:51:01.173534+00:00"}