{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:V26QLI5DQF67ZCNMHEEKUMENK3","short_pith_number":"pith:V26QLI5D","schema_version":"1.0","canonical_sha256":"aebd05a3a3817dfc89ac3908aa308d56faf3c6c3fb11da32b7b9c00b60dc10ff","source":{"kind":"arxiv","id":"2411.04996","version":2},"attestation_state":"computed","paper":{"title":"Mixture-of-Transformers: A Sparse and Scalable Architecture for Multi-Modal Foundation Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chunting Zhou, Gargi Ghosh, Liang Luo, Lili Yu, Luke Zettlemoyer, Mike Lewis, Ning Dong, Srinivasan Iyer, Weixin Liang, Wen-tau Yih, Xi Victoria Lin","submitted_at":"2024-11-07T18:59:06Z","abstract_excerpt":"The development of large language models (LLMs) has expanded to multi-modal systems capable of processing text, images, and speech within a unified framework. Training these models demands significantly larger datasets and computational resources compared to text-only LLMs. To address the scaling challenges, we introduce Mixture-of-Transformers (MoT), a sparse multi-modal transformer architecture that significantly reduces pretraining computational costs. MoT decouples non-embedding parameters of the model by modality -- including feed-forward networks, attention matrices, and layer normalizat"},"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":"2411.04996","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-07T18:59:06Z","cross_cats_sorted":[],"title_canon_sha256":"e079b34f86192301cb847ef4d01e5b1ef00de5dd55cddae41fba26115f8eb32b","abstract_canon_sha256":"63c23ae7820b209e313d61b2ebe13cee78cc1b6fa4f547e41b68ad0f8b5b475d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:43:43.371689Z","signature_b64":"hjGQXdKWLhAAl9necOepof4Ra/4FfiQU02YlgUP9A77XSqPfXU+HmbwjRKVb1gfHLP3UzYmxJEKLaljPg0CICQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aebd05a3a3817dfc89ac3908aa308d56faf3c6c3fb11da32b7b9c00b60dc10ff","last_reissued_at":"2026-05-18T02:43:43.370885Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:43:43.370885Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mixture-of-Transformers: A Sparse and Scalable Architecture for Multi-Modal Foundation Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chunting Zhou, Gargi Ghosh, Liang Luo, Lili Yu, Luke Zettlemoyer, Mike Lewis, Ning Dong, Srinivasan Iyer, Weixin Liang, Wen-tau Yih, Xi Victoria Lin","submitted_at":"2024-11-07T18:59:06Z","abstract_excerpt":"The development of large language models (LLMs) has expanded to multi-modal systems capable of processing text, images, and speech within a unified framework. Training these models demands significantly larger datasets and computational resources compared to text-only LLMs. To address the scaling challenges, we introduce Mixture-of-Transformers (MoT), a sparse multi-modal transformer architecture that significantly reduces pretraining computational costs. MoT decouples non-embedding parameters of the model by modality -- including feed-forward networks, attention matrices, and layer normalizat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.04996","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":""},"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":"2411.04996","created_at":"2026-05-18T02:43:43.371020+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.04996v2","created_at":"2026-05-18T02:43:43.371020+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.04996","created_at":"2026-05-18T02:43:43.371020+00:00"},{"alias_kind":"pith_short_12","alias_value":"V26QLI5DQF67","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"V26QLI5DQF67ZCNM","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"V26QLI5D","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":18,"internal_anchor_count":18,"sample":[{"citing_arxiv_id":"2507.15493","citing_title":"GR-3 Technical Report","ref_index":45,"is_internal_anchor":true},{"citing_arxiv_id":"2505.05472","citing_title":"Mogao: An Omni Foundation Model for Interleaved Multi-Modal Generation","ref_index":44,"is_internal_anchor":true},{"citing_arxiv_id":"2512.14234","citing_title":"ViBES: A Conversational Agent with Behaviorally-Intelligent 3D Virtual Body","ref_index":58,"is_internal_anchor":true},{"citing_arxiv_id":"2512.23421","citing_title":"DriveLaW:Unifying Planning and Video Generation in a Latent Driving World","ref_index":46,"is_internal_anchor":true},{"citing_arxiv_id":"2509.06951","citing_title":"F1: A Vision-Language-Action Model Bridging Understanding and Generation to Actions","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12622","citing_title":"Action Emergence from Streaming Intent","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12624","citing_title":"MindVLA-U1: VLA Beats VA with Unified Streaming Architecture for Autonomous Driving","ref_index":44,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12622","citing_title":"Action Emergence from Streaming Intent","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12624","citing_title":"MindVLA-U1: VLA Beats VA with Unified Streaming Architecture for Autonomous Driving","ref_index":44,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12500","citing_title":"SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture","ref_index":73,"is_internal_anchor":true},{"citing_arxiv_id":"2604.26341","citing_title":"SpatialFusion: Endowing Unified Image Generation with Intrinsic 3D Geometric Awareness","ref_index":26,"is_internal_anchor":true},{"citing_arxiv_id":"2604.24625","citing_title":"Meta-CoT: Enhancing Granularity and Generalization in Image Editing","ref_index":32,"is_internal_anchor":true},{"citing_arxiv_id":"2604.23272","citing_title":"Modular Sensory Stream for Integrating Physical Feedback in Vision-Language-Action Models","ref_index":8,"is_internal_anchor":true},{"citing_arxiv_id":"2605.05781","citing_title":"Steering Visual Generation in Unified Multimodal Models with Understanding Supervision","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"2604.19858","citing_title":"Wan-Image: Pushing the Boundaries of Generative Visual Intelligence","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2604.10949","citing_title":"Pseudo-Unification: Entropy Probing Reveals Divergent Information Patterns in Unified Multimodal Models","ref_index":34,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07794","citing_title":"NoiseGate: Learning Per-Latent Timestep Schedules as Information Gating in World Action Models","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2604.17706","citing_title":"OmniVLA-RL: A Vision-Language-Action Model with Spatial Understanding and Online RL","ref_index":49,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/V26QLI5DQF67ZCNMHEEKUMENK3","json":"https://pith.science/pith/V26QLI5DQF67ZCNMHEEKUMENK3.json","graph_json":"https://pith.science/api/pith-number/V26QLI5DQF67ZCNMHEEKUMENK3/graph.json","events_json":"https://pith.science/api/pith-number/V26QLI5DQF67ZCNMHEEKUMENK3/events.json","paper":"https://pith.science/paper/V26QLI5D"},"agent_actions":{"view_html":"https://pith.science/pith/V26QLI5DQF67ZCNMHEEKUMENK3","download_json":"https://pith.science/pith/V26QLI5DQF67ZCNMHEEKUMENK3.json","view_paper":"https://pith.science/paper/V26QLI5D","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.04996&json=true","fetch_graph":"https://pith.science/api/pith-number/V26QLI5DQF67ZCNMHEEKUMENK3/graph.json","fetch_events":"https://pith.science/api/pith-number/V26QLI5DQF67ZCNMHEEKUMENK3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/V26QLI5DQF67ZCNMHEEKUMENK3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/V26QLI5DQF67ZCNMHEEKUMENK3/action/storage_attestation","attest_author":"https://pith.science/pith/V26QLI5DQF67ZCNMHEEKUMENK3/action/author_attestation","sign_citation":"https://pith.science/pith/V26QLI5DQF67ZCNMHEEKUMENK3/action/citation_signature","submit_replication":"https://pith.science/pith/V26QLI5DQF67ZCNMHEEKUMENK3/action/replication_record"}},"created_at":"2026-05-18T02:43:43.371020+00:00","updated_at":"2026-05-18T02:43:43.371020+00:00"}