{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KA5QMZG7F2URFHNWZIZZTOYZ72","short_pith_number":"pith:KA5QMZG7","schema_version":"1.0","canonical_sha256":"503b0664df2ea9129db6ca3399bb19fe9a97eb97df65e97a94e46d2634b3191c","source":{"kind":"arxiv","id":"2603.09715","version":2},"attestation_state":"computed","paper":{"title":"Does the Question Really Matter? Training-Free Data Selection for Vision-Language SFT","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Huawen Shen, Peng Sun, Tianfan Fu, Yanbo Wang, Yi Ban, Yi Yang, Yuqiang Li","submitted_at":"2026-03-10T14:23:38Z","abstract_excerpt":"Visual instruction tuning is crucial for improving vision-language large models (VLLMs). However, many samples can be solved via linguistic patterns or common-sense shortcuts, without genuine cross-modal reasoning, limiting the effectiveness of multimodal learning. Prior data selection methods often rely on costly proxy model training and focus on difficulty or diversity, failing to capture a sample's true contribution to vision-language joint reasoning. In this paper, we propose CVS, a training-free data selection method based on the insight that, for high-quality multimodal samples, introduc"},"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":"2603.09715","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-03-10T14:23:38Z","cross_cats_sorted":[],"title_canon_sha256":"8d3e69c1d46344a871585c2fd4db63022ef96af7e21cf82166f0bfc23fd50daf","abstract_canon_sha256":"40d36ac2270ed6d40997a8df90cdcc3409620fce938478386fe085308e017baa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-11T01:09:32.765939Z","signature_b64":"zX3opUDq7NGTOYnSdin0jpUFAyUvfpJdlkvgre0ltOGHWk3WTjZJza9RPqlkHybpyDHIpDyg3Ske+4mWPnnyAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"503b0664df2ea9129db6ca3399bb19fe9a97eb97df65e97a94e46d2634b3191c","last_reissued_at":"2026-06-11T01:09:32.764898Z","signature_status":"signed_v1","first_computed_at":"2026-06-11T01:09:32.764898Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Does the Question Really Matter? Training-Free Data Selection for Vision-Language SFT","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Huawen Shen, Peng Sun, Tianfan Fu, Yanbo Wang, Yi Ban, Yi Yang, Yuqiang Li","submitted_at":"2026-03-10T14:23:38Z","abstract_excerpt":"Visual instruction tuning is crucial for improving vision-language large models (VLLMs). However, many samples can be solved via linguistic patterns or common-sense shortcuts, without genuine cross-modal reasoning, limiting the effectiveness of multimodal learning. Prior data selection methods often rely on costly proxy model training and focus on difficulty or diversity, failing to capture a sample's true contribution to vision-language joint reasoning. In this paper, we propose CVS, a training-free data selection method based on the insight that, for high-quality multimodal samples, introduc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.09715","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/2603.09715/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":"2603.09715","created_at":"2026-06-11T01:09:32.765055+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.09715v2","created_at":"2026-06-11T01:09:32.765055+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.09715","created_at":"2026-06-11T01:09:32.765055+00:00"},{"alias_kind":"pith_short_12","alias_value":"KA5QMZG7F2UR","created_at":"2026-06-11T01:09:32.765055+00:00"},{"alias_kind":"pith_short_16","alias_value":"KA5QMZG7F2URFHNW","created_at":"2026-06-11T01:09:32.765055+00:00"},{"alias_kind":"pith_short_8","alias_value":"KA5QMZG7","created_at":"2026-06-11T01:09:32.765055+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/KA5QMZG7F2URFHNWZIZZTOYZ72","json":"https://pith.science/pith/KA5QMZG7F2URFHNWZIZZTOYZ72.json","graph_json":"https://pith.science/api/pith-number/KA5QMZG7F2URFHNWZIZZTOYZ72/graph.json","events_json":"https://pith.science/api/pith-number/KA5QMZG7F2URFHNWZIZZTOYZ72/events.json","paper":"https://pith.science/paper/KA5QMZG7"},"agent_actions":{"view_html":"https://pith.science/pith/KA5QMZG7F2URFHNWZIZZTOYZ72","download_json":"https://pith.science/pith/KA5QMZG7F2URFHNWZIZZTOYZ72.json","view_paper":"https://pith.science/paper/KA5QMZG7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.09715&json=true","fetch_graph":"https://pith.science/api/pith-number/KA5QMZG7F2URFHNWZIZZTOYZ72/graph.json","fetch_events":"https://pith.science/api/pith-number/KA5QMZG7F2URFHNWZIZZTOYZ72/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KA5QMZG7F2URFHNWZIZZTOYZ72/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KA5QMZG7F2URFHNWZIZZTOYZ72/action/storage_attestation","attest_author":"https://pith.science/pith/KA5QMZG7F2URFHNWZIZZTOYZ72/action/author_attestation","sign_citation":"https://pith.science/pith/KA5QMZG7F2URFHNWZIZZTOYZ72/action/citation_signature","submit_replication":"https://pith.science/pith/KA5QMZG7F2URFHNWZIZZTOYZ72/action/replication_record"}},"created_at":"2026-06-11T01:09:32.765055+00:00","updated_at":"2026-06-11T01:09:32.765055+00:00"}