{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:R24PPKZ5T7XMD642ZIA7OMLVYY","short_pith_number":"pith:R24PPKZ5","schema_version":"1.0","canonical_sha256":"8eb8f7ab3d9feec1fb9aca01f73175c60ef562b03e77b1bfe8c10edabfc2e360","source":{"kind":"arxiv","id":"2506.17901","version":1},"attestation_state":"computed","paper":{"title":"PostAlign: Multimodal Grounding as a Corrective Lens for MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jian Wu, Jindong Gu, Philip Torr, Yang Zhang, Yixuan Wu","submitted_at":"2025-06-22T05:11:46Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) excel in vision-language tasks, such as image captioning and visual question answering. However, they often suffer from over-reliance on spurious correlations, primarily due to linguistic priors that distract the model from leveraging actual visual information. To address these issues, we introduce MMGrounded-PostAlign, a post-multimodal alignment framework designed to enhance the visual understanding capabilities and mitigate the hallucinations of MLLMs. Our framework incorporates a multimodal grounding module for both visual grounding, which identifie"},"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":"2506.17901","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-06-22T05:11:46Z","cross_cats_sorted":[],"title_canon_sha256":"0e82280c54cfd192c6aae9c52a83bc6cfaa7497adaa69293f69ee97a4f90aa91","abstract_canon_sha256":"a9ce105960cab0257ba50a2a1696155cc7a9bb181dfa678711f72e5af4a4df27"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:25:25.156995Z","signature_b64":"IRVLG2fUKpsmcH0OCoi7Bv3LYQYDk1cIPJcWquKrdujG2meq63EWIBsJ0SmhS0f5I45/EG0dQqBbRXuPJGoPBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8eb8f7ab3d9feec1fb9aca01f73175c60ef562b03e77b1bfe8c10edabfc2e360","last_reissued_at":"2026-07-05T11:25:25.156593Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:25:25.156593Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PostAlign: Multimodal Grounding as a Corrective Lens for MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jian Wu, Jindong Gu, Philip Torr, Yang Zhang, Yixuan Wu","submitted_at":"2025-06-22T05:11:46Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) excel in vision-language tasks, such as image captioning and visual question answering. However, they often suffer from over-reliance on spurious correlations, primarily due to linguistic priors that distract the model from leveraging actual visual information. To address these issues, we introduce MMGrounded-PostAlign, a post-multimodal alignment framework designed to enhance the visual understanding capabilities and mitigate the hallucinations of MLLMs. Our framework incorporates a multimodal grounding module for both visual grounding, which identifie"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.17901","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/2506.17901/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":"2506.17901","created_at":"2026-07-05T11:25:25.156651+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.17901v1","created_at":"2026-07-05T11:25:25.156651+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.17901","created_at":"2026-07-05T11:25:25.156651+00:00"},{"alias_kind":"pith_short_12","alias_value":"R24PPKZ5T7XM","created_at":"2026-07-05T11:25:25.156651+00:00"},{"alias_kind":"pith_short_16","alias_value":"R24PPKZ5T7XMD642","created_at":"2026-07-05T11:25:25.156651+00:00"},{"alias_kind":"pith_short_8","alias_value":"R24PPKZ5","created_at":"2026-07-05T11:25:25.156651+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/R24PPKZ5T7XMD642ZIA7OMLVYY","json":"https://pith.science/pith/R24PPKZ5T7XMD642ZIA7OMLVYY.json","graph_json":"https://pith.science/api/pith-number/R24PPKZ5T7XMD642ZIA7OMLVYY/graph.json","events_json":"https://pith.science/api/pith-number/R24PPKZ5T7XMD642ZIA7OMLVYY/events.json","paper":"https://pith.science/paper/R24PPKZ5"},"agent_actions":{"view_html":"https://pith.science/pith/R24PPKZ5T7XMD642ZIA7OMLVYY","download_json":"https://pith.science/pith/R24PPKZ5T7XMD642ZIA7OMLVYY.json","view_paper":"https://pith.science/paper/R24PPKZ5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.17901&json=true","fetch_graph":"https://pith.science/api/pith-number/R24PPKZ5T7XMD642ZIA7OMLVYY/graph.json","fetch_events":"https://pith.science/api/pith-number/R24PPKZ5T7XMD642ZIA7OMLVYY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/R24PPKZ5T7XMD642ZIA7OMLVYY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/R24PPKZ5T7XMD642ZIA7OMLVYY/action/storage_attestation","attest_author":"https://pith.science/pith/R24PPKZ5T7XMD642ZIA7OMLVYY/action/author_attestation","sign_citation":"https://pith.science/pith/R24PPKZ5T7XMD642ZIA7OMLVYY/action/citation_signature","submit_replication":"https://pith.science/pith/R24PPKZ5T7XMD642ZIA7OMLVYY/action/replication_record"}},"created_at":"2026-07-05T11:25:25.156651+00:00","updated_at":"2026-07-05T11:25:25.156651+00:00"}