{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BY5ENMHMU33RX7RAVCCVXGVSZU","short_pith_number":"pith:BY5ENMHM","schema_version":"1.0","canonical_sha256":"0e3a46b0eca6f71bfe20a8855b9ab2cd08d9a75dab9d31de97fddc30d9ced46c","source":{"kind":"arxiv","id":"2412.14171","version":2},"attestation_state":"computed","paper":{"title":"Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Anjali W. Gupta, Jihan Yang, Li Fei-Fei, Rilyn Han, Saining Xie, Shusheng Yang","submitted_at":"2024-12-18T18:59:54Z","abstract_excerpt":"Humans possess the visual-spatial intelligence to remember spaces from sequential visual observations. However, can Multimodal Large Language Models (MLLMs) trained on million-scale video datasets also ``think in space'' from videos? We present a novel video-based visual-spatial intelligence benchmark (VSI-Bench) of over 5,000 question-answer pairs, and find that MLLMs exhibit competitive - though subhuman - visual-spatial intelligence. We probe models to express how they think in space both linguistically and visually and find that while spatial reasoning capabilities remain the primary bottl"},"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":"2412.14171","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-12-18T18:59:54Z","cross_cats_sorted":[],"title_canon_sha256":"3139208067f1984a077d0183513fa153b33c1a95940c0febe9f9c6dfda7e9b77","abstract_canon_sha256":"c0b08d0db958eccefed1a4a70839bbb28797d599f952e5ab21f6c7ba12a54183"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-22T09:22:37.130500Z","signature_b64":"sQ1lFGdzdkOPE6DverQeneKV5FCgvfOHMEELE7jCd41oQZOv97WCEyhfQS1B7sYm+ssNC0xvivInk3myZBgyBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0e3a46b0eca6f71bfe20a8855b9ab2cd08d9a75dab9d31de97fddc30d9ced46c","last_reissued_at":"2026-05-22T09:22:37.127517Z","signature_status":"signed_v1","first_computed_at":"2026-05-22T09:22:37.127517Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Anjali W. Gupta, Jihan Yang, Li Fei-Fei, Rilyn Han, Saining Xie, Shusheng Yang","submitted_at":"2024-12-18T18:59:54Z","abstract_excerpt":"Humans possess the visual-spatial intelligence to remember spaces from sequential visual observations. However, can Multimodal Large Language Models (MLLMs) trained on million-scale video datasets also ``think in space'' from videos? We present a novel video-based visual-spatial intelligence benchmark (VSI-Bench) of over 5,000 question-answer pairs, and find that MLLMs exhibit competitive - though subhuman - visual-spatial intelligence. We probe models to express how they think in space both linguistically and visually and find that while spatial reasoning capabilities remain the primary bottl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.14171","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/2412.14171/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":"2412.14171","created_at":"2026-05-22T09:22:37.127684+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.14171v2","created_at":"2026-05-22T09:22:37.127684+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.14171","created_at":"2026-05-22T09:22:37.127684+00:00"},{"alias_kind":"pith_short_12","alias_value":"BY5ENMHMU33R","created_at":"2026-05-22T09:22:37.127684+00:00"},{"alias_kind":"pith_short_16","alias_value":"BY5ENMHMU33RX7RA","created_at":"2026-05-22T09:22:37.127684+00:00"},{"alias_kind":"pith_short_8","alias_value":"BY5ENMHM","created_at":"2026-05-22T09:22:37.127684+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":24,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.21642","citing_title":"Ablate-to-Validate: Are Vision-Language Models Really Using Continuous Thought Tokens?","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"2605.21625","citing_title":"Flat-Pack Bench: Evaluating Spatio-Temporal Understanding in Large Vision-Language Models through Furniture Assembly","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22536","citing_title":"SpaceDG: Benchmarking Spatial Intelligence under Visual Degradation","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2505.23678","citing_title":"Grounded Reinforcement Learning for Visual Reasoning","ref_index":76,"is_internal_anchor":false},{"citing_arxiv_id":"2505.23747","citing_title":"Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15777","citing_title":"SaaS-Bench: Can Computer-Use Agents Leverage Real-World SaaS to Solve Professional Workflows?","ref_index":88,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16899","citing_title":"LASAR: Towards Spatio-temporal Reasoning with Latent Cognitive Map","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2505.20279","citing_title":"VLM-3R: Vision-Language Models Augmented with Instruction-Aligned 3D Reconstruction","ref_index":77,"is_internal_anchor":false},{"citing_arxiv_id":"2506.06856","citing_title":"Vision-EKIPL: External Knowledge-Infused Policy Learning for Visual Reasoning","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2508.13998","citing_title":"Embodied-R1: Reinforced Embodied Reasoning for General Robotic Manipulation","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2505.21374","citing_title":"Video-Holmes: Can MLLM Think Like Holmes for Complex Video Reasoning?","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2506.09965","citing_title":"Reinforcing Spatial Reasoning in Vision-Language Models with Interwoven Thinking and Visual Drawing","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2505.23747","citing_title":"Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2503.12605","citing_title":"Multimodal Chain-of-Thought Reasoning: A Comprehensive Survey","ref_index":169,"is_internal_anchor":false},{"citing_arxiv_id":"2504.01805","citing_title":"SpaceR: Reinforcing MLLMs in Video Spatial Reasoning","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2503.21776","citing_title":"Video-R1: Reinforcing Video Reasoning in MLLMs","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2501.15830","citing_title":"SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Model","ref_index":67,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09449","citing_title":"SpaceMind++: Toward Allocentric Cognitive Maps for Spatially Grounded Video MLLMs","ref_index":67,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20473","citing_title":"Video-ToC: Video Tree-of-Cue Reasoning","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09535","citing_title":"EgoTL: Egocentric Think-Aloud Chains for Long-Horizon Tasks","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09167","citing_title":"MAG-3D: Multi-Agent Grounded Reasoning for 3D Understanding","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2504.10479","citing_title":"InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models","ref_index":135,"is_internal_anchor":false},{"citing_arxiv_id":"2508.18265","citing_title":"InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency","ref_index":162,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16893","citing_title":"EasyVideoR1: Easier RL for Video Understanding","ref_index":46,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BY5ENMHMU33RX7RAVCCVXGVSZU","json":"https://pith.science/pith/BY5ENMHMU33RX7RAVCCVXGVSZU.json","graph_json":"https://pith.science/api/pith-number/BY5ENMHMU33RX7RAVCCVXGVSZU/graph.json","events_json":"https://pith.science/api/pith-number/BY5ENMHMU33RX7RAVCCVXGVSZU/events.json","paper":"https://pith.science/paper/BY5ENMHM"},"agent_actions":{"view_html":"https://pith.science/pith/BY5ENMHMU33RX7RAVCCVXGVSZU","download_json":"https://pith.science/pith/BY5ENMHMU33RX7RAVCCVXGVSZU.json","view_paper":"https://pith.science/paper/BY5ENMHM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.14171&json=true","fetch_graph":"https://pith.science/api/pith-number/BY5ENMHMU33RX7RAVCCVXGVSZU/graph.json","fetch_events":"https://pith.science/api/pith-number/BY5ENMHMU33RX7RAVCCVXGVSZU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BY5ENMHMU33RX7RAVCCVXGVSZU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BY5ENMHMU33RX7RAVCCVXGVSZU/action/storage_attestation","attest_author":"https://pith.science/pith/BY5ENMHMU33RX7RAVCCVXGVSZU/action/author_attestation","sign_citation":"https://pith.science/pith/BY5ENMHMU33RX7RAVCCVXGVSZU/action/citation_signature","submit_replication":"https://pith.science/pith/BY5ENMHMU33RX7RAVCCVXGVSZU/action/replication_record"}},"created_at":"2026-05-22T09:22:37.127684+00:00","updated_at":"2026-05-22T09:22:37.127684+00:00"}