{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:QFQWN44TNS5CUJPHFE6PAGGWNM","short_pith_number":"pith:QFQWN44T","schema_version":"1.0","canonical_sha256":"816166f3936cba2a25e7293cf018d66b041c13d14c6edcd30bb7686fb18cfb0c","source":{"kind":"arxiv","id":"2505.14640","version":1},"attestation_state":"computed","paper":{"title":"VideoEval-Pro: Robust and Realistic Long Video Understanding Evaluation","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ge Zhang, Ping Nie, Weiming Ren, Wenhu Chen, Wentao Ma, Yiming Jia, Zhuofeng Li","submitted_at":"2025-05-20T17:26:32Z","abstract_excerpt":"Large multimodal models (LMMs) have recently emerged as a powerful tool for long video understanding (LVU), prompting the development of standardized LVU benchmarks to evaluate their performance. However, our investigation reveals a rather sober lesson for existing LVU benchmarks. First, most existing benchmarks rely heavily on multiple-choice questions (MCQs), whose evaluation results are inflated due to the possibility of guessing the correct answer; Second, a significant portion of questions in these benchmarks have strong priors to allow models to answer directly without even reading the i"},"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":"2505.14640","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-20T17:26:32Z","cross_cats_sorted":[],"title_canon_sha256":"5022d687ef4d596d506f79047b49527ed014708909a3507703e9bffbbe2e15c9","abstract_canon_sha256":"b8e7585cc723611c02d776132ec6177a1fe812a79e323542a9cfc28f721cbe91"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:06:16.894577Z","signature_b64":"qz+dF7lL6995F2SANkn3jb4i4XA3nVcCaH4QOAU8Poq1AJFx4yFxbnCcrxl5D/shG2NaPrb+FPcFp5g81NLXBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"816166f3936cba2a25e7293cf018d66b041c13d14c6edcd30bb7686fb18cfb0c","last_reissued_at":"2026-07-05T11:06:16.894054Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:06:16.894054Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VideoEval-Pro: Robust and Realistic Long Video Understanding Evaluation","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ge Zhang, Ping Nie, Weiming Ren, Wenhu Chen, Wentao Ma, Yiming Jia, Zhuofeng Li","submitted_at":"2025-05-20T17:26:32Z","abstract_excerpt":"Large multimodal models (LMMs) have recently emerged as a powerful tool for long video understanding (LVU), prompting the development of standardized LVU benchmarks to evaluate their performance. However, our investigation reveals a rather sober lesson for existing LVU benchmarks. First, most existing benchmarks rely heavily on multiple-choice questions (MCQs), whose evaluation results are inflated due to the possibility of guessing the correct answer; Second, a significant portion of questions in these benchmarks have strong priors to allow models to answer directly without even reading the i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.14640","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/2505.14640/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":"2505.14640","created_at":"2026-07-05T11:06:16.894122+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.14640v1","created_at":"2026-07-05T11:06:16.894122+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.14640","created_at":"2026-07-05T11:06:16.894122+00:00"},{"alias_kind":"pith_short_12","alias_value":"QFQWN44TNS5C","created_at":"2026-07-05T11:06:16.894122+00:00"},{"alias_kind":"pith_short_16","alias_value":"QFQWN44TNS5CUJPH","created_at":"2026-07-05T11:06:16.894122+00:00"},{"alias_kind":"pith_short_8","alias_value":"QFQWN44T","created_at":"2026-07-05T11:06:16.894122+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.00248","citing_title":"Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2601.10611","citing_title":"Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding","ref_index":98,"is_internal_anchor":false},{"citing_arxiv_id":"2602.22779","citing_title":"TrajTok: Learning Trajectory Tokens enables better Video Understanding","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10434","citing_title":"WorldReasonBench: Human-Aligned Stress Testing of Video Generators as Future World-State Predictors","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QFQWN44TNS5CUJPHFE6PAGGWNM","json":"https://pith.science/pith/QFQWN44TNS5CUJPHFE6PAGGWNM.json","graph_json":"https://pith.science/api/pith-number/QFQWN44TNS5CUJPHFE6PAGGWNM/graph.json","events_json":"https://pith.science/api/pith-number/QFQWN44TNS5CUJPHFE6PAGGWNM/events.json","paper":"https://pith.science/paper/QFQWN44T"},"agent_actions":{"view_html":"https://pith.science/pith/QFQWN44TNS5CUJPHFE6PAGGWNM","download_json":"https://pith.science/pith/QFQWN44TNS5CUJPHFE6PAGGWNM.json","view_paper":"https://pith.science/paper/QFQWN44T","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.14640&json=true","fetch_graph":"https://pith.science/api/pith-number/QFQWN44TNS5CUJPHFE6PAGGWNM/graph.json","fetch_events":"https://pith.science/api/pith-number/QFQWN44TNS5CUJPHFE6PAGGWNM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QFQWN44TNS5CUJPHFE6PAGGWNM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QFQWN44TNS5CUJPHFE6PAGGWNM/action/storage_attestation","attest_author":"https://pith.science/pith/QFQWN44TNS5CUJPHFE6PAGGWNM/action/author_attestation","sign_citation":"https://pith.science/pith/QFQWN44TNS5CUJPHFE6PAGGWNM/action/citation_signature","submit_replication":"https://pith.science/pith/QFQWN44TNS5CUJPHFE6PAGGWNM/action/replication_record"}},"created_at":"2026-07-05T11:06:16.894122+00:00","updated_at":"2026-07-05T11:06:16.894122+00:00"}