{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:SGUFOMNHKQL4IFMWYUVCLHHQH6","short_pith_number":"pith:SGUFOMNH","canonical_record":{"source":{"id":"2503.07536","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-10T17:04:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"43e5c65607dd7bda3cdd188b43717b8d938f768ed61b7ba0f73373e48bba2933","abstract_canon_sha256":"153a64ed5dc6d65cf29c723c8da3f7bac130de117b7afe63774316ea84075c24"},"schema_version":"1.0"},"canonical_sha256":"91a85731a75417c41596c52a259cf03f90f2493edb555761d42ec860aaa2051f","source":{"kind":"arxiv","id":"2503.07536","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.07536","created_at":"2026-05-17T23:38:47Z"},{"alias_kind":"arxiv_version","alias_value":"2503.07536v2","created_at":"2026-05-17T23:38:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.07536","created_at":"2026-05-17T23:38:47Z"},{"alias_kind":"pith_short_12","alias_value":"SGUFOMNHKQL4","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"SGUFOMNHKQL4IFMW","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"SGUFOMNH","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:SGUFOMNHKQL4IFMWYUVCLHHQH6","target":"record","payload":{"canonical_record":{"source":{"id":"2503.07536","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-10T17:04:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"43e5c65607dd7bda3cdd188b43717b8d938f768ed61b7ba0f73373e48bba2933","abstract_canon_sha256":"153a64ed5dc6d65cf29c723c8da3f7bac130de117b7afe63774316ea84075c24"},"schema_version":"1.0"},"canonical_sha256":"91a85731a75417c41596c52a259cf03f90f2493edb555761d42ec860aaa2051f","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:47.472006Z","signature_b64":"nuf+6QGYCQHN99qO4zCtQu71ndGx8bfizMXlsVpRIU9b522Hv1M6ERd9kFH5/ynZUtX+shKNeKTrWicpwmahDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"91a85731a75417c41596c52a259cf03f90f2493edb555761d42ec860aaa2051f","last_reissued_at":"2026-05-17T23:38:47.471538Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:47.471538Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2503.07536","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rg1ym89tQAyEr5TXt1rK+1xhJk7iX6qFRl8ar3gNAxufjwqS4UpyMC86YZz2Tt4tRZJU4sL2TnOGrbEDy3uICA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T23:45:22.950575Z"},"content_sha256":"3050734ff785efa280e7dd156dc29be88ab79aa0ff77e96cc8bc0cc998d20d59","schema_version":"1.0","event_id":"sha256:3050734ff785efa280e7dd156dc29be88ab79aa0ff77e96cc8bc0cc998d20d59"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:SGUFOMNHKQL4IFMWYUVCLHHQH6","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Gongrui Zhang, Jie Liu, Kai Yang, Miaosen Zhang, Qipeng Zhu, Xin Geng, Xingzhong Xu, Xu Yang, Yingzhe Peng, Zhiyuan You","submitted_at":"2025-03-10T17:04:14Z","abstract_excerpt":"Enhancing reasoning in Large Multimodal Models (LMMs) faces unique challenges from the complex interplay between visual perception and logical reasoning, particularly in compact 3B-parameter architectures where architectural constraints limit reasoning capacity and modality alignment.\n  While rule-based reinforcement learning (RL) excels in text-only domains, its multimodal extension confronts two critical barriers: (1) data limitations due to ambiguous answers and scarce complex reasoning examples, and (2) degraded foundational reasoning induced by multimodal pretraining. To address these cha"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"text-based reasoning enhancement enables effective multimodal generalization, offering a data-efficient paradigm that bypasses costly high-quality multimodal training data.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That reasoning skills strengthened via text-only rule-based RL transfer to multimodal inputs without substantial interference from visual perception components or degradation of the foundational reasoning.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A two-stage RL framework first boosts text reasoning in 3B LMMs then adapts it to multimodal inputs, producing modest benchmark gains of 4.5-4.8%.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"86c3f6d99d27df1b124bf81641972e85a603eca4101442a220dc3281f3f85af3"},"source":{"id":"2503.07536","kind":"arxiv","version":2},"verdict":{"id":"9c7ebaba-1a3d-46f4-a4f2-4d7f1402af82","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T15:11:35.828514Z","strongest_claim":"text-based reasoning enhancement enables effective multimodal generalization, offering a data-efficient paradigm that bypasses costly high-quality multimodal training data.","one_line_summary":"A two-stage RL framework first boosts text reasoning in 3B LMMs then adapts it to multimodal inputs, producing modest benchmark gains of 4.5-4.8%.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That reasoning skills strengthened via text-only rule-based RL transfer to multimodal inputs without substantial interference from visual perception components or degradation of the foundational reasoning.","pith_extraction_headline":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data."},"references":{"count":117,"sample":[{"doi":"","year":null,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","ref_index":1,"cited_arxiv_id":"2303.08774","is_internal_anchor":true},{"doi":"","year":2015,"title":"Lawrence Zitnick, Devi Parikh, and Dhruv Ba- tra","work_id":"7945da46-3915-4108-853a-ab77b7e0cd3c","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Flamingo: A visual language model for few-shot learning","work_id":"1863a034-835c-4304-b6c6-32734dcf9e1a","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Qwen Technical Report","work_id":"bb1fd52f-6b2f-437c-9516-37bdf6eb9be8","ref_index":4,"cited_arxiv_id":"2309.16609","is_internal_anchor":true},{"doi":"","year":2023,"title":"Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond","work_id":"cbc2bb21-b6bb-46c0-80bf-107e195ffe10","ref_index":5,"cited_arxiv_id":"2308.12966","is_internal_anchor":true}],"resolved_work":117,"snapshot_sha256":"19d1ebaaa5411c244a51a482d94aacad3b045979c3b0e1835cfd65922d00882d","internal_anchors":31},"formal_canon":{"evidence_count":2,"snapshot_sha256":"1f56fb3c6ceb513d5107c257c4adc2804883b4df9c9db298f66864c895bfa6d4"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"9c7ebaba-1a3d-46f4-a4f2-4d7f1402af82"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FQSngLghLAYbLp1lnXNeIl9e11ILnBUvE2Fi/KAkbVPToa4qJXYBsp9wNlGsTQzLoKULr1qXfNAk/jwS0VgBCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T23:45:22.951238Z"},"content_sha256":"c14fa2237a3e38356abffadd4a8e23ab757b4753d2968d73e89efbd081b7bb01","schema_version":"1.0","event_id":"sha256:c14fa2237a3e38356abffadd4a8e23ab757b4753d2968d73e89efbd081b7bb01"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/bundle.json","state_url":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-25T23:45:22Z","links":{"resolver":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6","bundle":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/bundle.json","state":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:SGUFOMNHKQL4IFMWYUVCLHHQH6","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"153a64ed5dc6d65cf29c723c8da3f7bac130de117b7afe63774316ea84075c24","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-10T17:04:14Z","title_canon_sha256":"43e5c65607dd7bda3cdd188b43717b8d938f768ed61b7ba0f73373e48bba2933"},"schema_version":"1.0","source":{"id":"2503.07536","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.07536","created_at":"2026-05-17T23:38:47Z"},{"alias_kind":"arxiv_version","alias_value":"2503.07536v2","created_at":"2026-05-17T23:38:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.07536","created_at":"2026-05-17T23:38:47Z"},{"alias_kind":"pith_short_12","alias_value":"SGUFOMNHKQL4","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"SGUFOMNHKQL4IFMW","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"SGUFOMNH","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:c14fa2237a3e38356abffadd4a8e23ab757b4753d2968d73e89efbd081b7bb01","target":"graph","created_at":"2026-05-17T23:38:47Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"text-based reasoning enhancement enables effective multimodal generalization, offering a data-efficient paradigm that bypasses costly high-quality multimodal training data."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That reasoning skills strengthened via text-only rule-based RL transfer to multimodal inputs without substantial interference from visual perception components or degradation of the foundational reasoning."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"A two-stage RL framework first boosts text reasoning in 3B LMMs then adapts it to multimodal inputs, producing modest benchmark gains of 4.5-4.8%."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data."}],"snapshot_sha256":"86c3f6d99d27df1b124bf81641972e85a603eca4101442a220dc3281f3f85af3"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"1f56fb3c6ceb513d5107c257c4adc2804883b4df9c9db298f66864c895bfa6d4"},"paper":{"abstract_excerpt":"Enhancing reasoning in Large Multimodal Models (LMMs) faces unique challenges from the complex interplay between visual perception and logical reasoning, particularly in compact 3B-parameter architectures where architectural constraints limit reasoning capacity and modality alignment.\n  While rule-based reinforcement learning (RL) excels in text-only domains, its multimodal extension confronts two critical barriers: (1) data limitations due to ambiguous answers and scarce complex reasoning examples, and (2) degraded foundational reasoning induced by multimodal pretraining. To address these cha","authors_text":"Gongrui Zhang, Jie Liu, Kai Yang, Miaosen Zhang, Qipeng Zhu, Xin Geng, Xingzhong Xu, Xu Yang, Yingzhe Peng, Zhiyuan You","cross_cats":["cs.AI"],"headline":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-10T17:04:14Z","title":"LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL"},"references":{"count":117,"internal_anchors":31,"resolved_work":117,"sample":[{"cited_arxiv_id":"2303.08774","doi":"","is_internal_anchor":true,"ref_index":1,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Lawrence Zitnick, Devi Parikh, and Dhruv Ba- tra","work_id":"7945da46-3915-4108-853a-ab77b7e0cd3c","year":2015},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Flamingo: A visual language model for few-shot learning","work_id":"1863a034-835c-4304-b6c6-32734dcf9e1a","year":null},{"cited_arxiv_id":"2309.16609","doi":"","is_internal_anchor":true,"ref_index":4,"title":"Qwen Technical Report","work_id":"bb1fd52f-6b2f-437c-9516-37bdf6eb9be8","year":2023},{"cited_arxiv_id":"2308.12966","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond","work_id":"cbc2bb21-b6bb-46c0-80bf-107e195ffe10","year":2023}],"snapshot_sha256":"19d1ebaaa5411c244a51a482d94aacad3b045979c3b0e1835cfd65922d00882d"},"source":{"id":"2503.07536","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-16T15:11:35.828514Z","id":"9c7ebaba-1a3d-46f4-a4f2-4d7f1402af82","model_set":{"reader":"grok-4.3"},"one_line_summary":"A two-stage RL framework first boosts text reasoning in 3B LMMs then adapts it to multimodal inputs, producing modest benchmark gains of 4.5-4.8%.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data.","strongest_claim":"text-based reasoning enhancement enables effective multimodal generalization, offering a data-efficient paradigm that bypasses costly high-quality multimodal training data.","weakest_assumption":"That reasoning skills strengthened via text-only rule-based RL transfer to multimodal inputs without substantial interference from visual perception components or degradation of the foundational reasoning."}},"verdict_id":"9c7ebaba-1a3d-46f4-a4f2-4d7f1402af82"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:3050734ff785efa280e7dd156dc29be88ab79aa0ff77e96cc8bc0cc998d20d59","target":"record","created_at":"2026-05-17T23:38:47Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"153a64ed5dc6d65cf29c723c8da3f7bac130de117b7afe63774316ea84075c24","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-10T17:04:14Z","title_canon_sha256":"43e5c65607dd7bda3cdd188b43717b8d938f768ed61b7ba0f73373e48bba2933"},"schema_version":"1.0","source":{"id":"2503.07536","kind":"arxiv","version":2}},"canonical_sha256":"91a85731a75417c41596c52a259cf03f90f2493edb555761d42ec860aaa2051f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"91a85731a75417c41596c52a259cf03f90f2493edb555761d42ec860aaa2051f","first_computed_at":"2026-05-17T23:38:47.471538Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:47.471538Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nuf+6QGYCQHN99qO4zCtQu71ndGx8bfizMXlsVpRIU9b522Hv1M6ERd9kFH5/ynZUtX+shKNeKTrWicpwmahDg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:47.472006Z","signed_message":"canonical_sha256_bytes"},"source_id":"2503.07536","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3050734ff785efa280e7dd156dc29be88ab79aa0ff77e96cc8bc0cc998d20d59","sha256:c14fa2237a3e38356abffadd4a8e23ab757b4753d2968d73e89efbd081b7bb01"],"state_sha256":"f487fd8eafce1b5aa4b04b2da3153f9ff1366d6e19fc914f2f7cfe95e09691a4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wyvWswVp6PI8Tixt2YtEhfXfW3fhx4xTGwIpM8oPX/iE9js0CeN34xWgMCcd1hEu62ew8rYbpFNKtYkQaGHxBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-25T23:45:22.955919Z","bundle_sha256":"11bb4c06d3824d60f23047718eec43603f48eb719a4b3b554818b8bfc86ca614"}}