{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:Y27RRRO4USDHW5FFDIYVXKKMZU","short_pith_number":"pith:Y27RRRO4","schema_version":"1.0","canonical_sha256":"c6bf18c5dca4867b74a51a315ba94ccd1ca608ee5e8cfefbef669d42f175c0fc","source":{"kind":"arxiv","id":"2605.20795","version":1},"attestation_state":"computed","paper":{"title":"What Semantics Survive the Connector? Diagnosing VLM-to-DiT Alignment in Video Editing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chao Wen, Chengming Xu, Hangyu Lin, Jiangning Zhang, Jianxiong Gao, Xiaobin Hu, Yanwei Fu","submitted_at":"2026-05-20T06:42:15Z","abstract_excerpt":"Flow matching based video generative models have been increasingly relying on prepended Vision-Language Models (VLMs) to handle complex, instruction-based video editing. The prevailing assumption underlying this paradigm is that a connector module can seamlessly align the VLM's rich multi-modal reasoning with the original text embedding space of DiTs. However, we hypothesize that this alignment acts as a severe semantic bottleneck, degrading fine-grained structural variables. Verifying this is challenging, as end-to-end evaluations conflate alignment failures with generation errors, and natura"},"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":"2605.20795","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-20T06:42:15Z","cross_cats_sorted":[],"title_canon_sha256":"36c8f3a99d52f42bf2a371af83b6704cdaccd76d8b53ea72f1d69687f98c9f0a","abstract_canon_sha256":"4cb4d6c16ae4d4bb89d378cfb09df6fde05346d6a9545cc589d7796e0c26a7e0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:04:54.893196Z","signature_b64":"GvP7ciLAv4dVYFTGZ34AwtXynXzMw7LDkfvzY4qKnsZWjrCluxamRoOXKxwCRiYhhRaNy5+7TJ4wRjXYjh5/AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c6bf18c5dca4867b74a51a315ba94ccd1ca608ee5e8cfefbef669d42f175c0fc","last_reissued_at":"2026-05-21T01:04:54.892219Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:04:54.892219Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"What Semantics Survive the Connector? Diagnosing VLM-to-DiT Alignment in Video Editing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chao Wen, Chengming Xu, Hangyu Lin, Jiangning Zhang, Jianxiong Gao, Xiaobin Hu, Yanwei Fu","submitted_at":"2026-05-20T06:42:15Z","abstract_excerpt":"Flow matching based video generative models have been increasingly relying on prepended Vision-Language Models (VLMs) to handle complex, instruction-based video editing. The prevailing assumption underlying this paradigm is that a connector module can seamlessly align the VLM's rich multi-modal reasoning with the original text embedding space of DiTs. However, we hypothesize that this alignment acts as a severe semantic bottleneck, degrading fine-grained structural variables. Verifying this is challenging, as end-to-end evaluations conflate alignment failures with generation errors, and natura"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20795","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/2605.20795/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":"2605.20795","created_at":"2026-05-21T01:04:54.892379+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20795v1","created_at":"2026-05-21T01:04:54.892379+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20795","created_at":"2026-05-21T01:04:54.892379+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y27RRRO4USDH","created_at":"2026-05-21T01:04:54.892379+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y27RRRO4USDHW5FF","created_at":"2026-05-21T01:04:54.892379+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y27RRRO4","created_at":"2026-05-21T01:04:54.892379+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/Y27RRRO4USDHW5FFDIYVXKKMZU","json":"https://pith.science/pith/Y27RRRO4USDHW5FFDIYVXKKMZU.json","graph_json":"https://pith.science/api/pith-number/Y27RRRO4USDHW5FFDIYVXKKMZU/graph.json","events_json":"https://pith.science/api/pith-number/Y27RRRO4USDHW5FFDIYVXKKMZU/events.json","paper":"https://pith.science/paper/Y27RRRO4"},"agent_actions":{"view_html":"https://pith.science/pith/Y27RRRO4USDHW5FFDIYVXKKMZU","download_json":"https://pith.science/pith/Y27RRRO4USDHW5FFDIYVXKKMZU.json","view_paper":"https://pith.science/paper/Y27RRRO4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20795&json=true","fetch_graph":"https://pith.science/api/pith-number/Y27RRRO4USDHW5FFDIYVXKKMZU/graph.json","fetch_events":"https://pith.science/api/pith-number/Y27RRRO4USDHW5FFDIYVXKKMZU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y27RRRO4USDHW5FFDIYVXKKMZU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y27RRRO4USDHW5FFDIYVXKKMZU/action/storage_attestation","attest_author":"https://pith.science/pith/Y27RRRO4USDHW5FFDIYVXKKMZU/action/author_attestation","sign_citation":"https://pith.science/pith/Y27RRRO4USDHW5FFDIYVXKKMZU/action/citation_signature","submit_replication":"https://pith.science/pith/Y27RRRO4USDHW5FFDIYVXKKMZU/action/replication_record"}},"created_at":"2026-05-21T01:04:54.892379+00:00","updated_at":"2026-05-21T01:04:54.892379+00:00"}