{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:M2OT5MWAUDNN6NLLCEHNG4CN36","short_pith_number":"pith:M2OT5MWA","schema_version":"1.0","canonical_sha256":"669d3eb2c0a0dadf356b110ed3704ddfb85e0c3f3d1f6296844826b6a4db0de9","source":{"kind":"arxiv","id":"2503.01329","version":2},"attestation_state":"computed","paper":{"title":"Neural ODE Transformers: Analyzing Internal Dynamics and Adaptive Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Anh Tong, Cheongwoong Kang, David Hall, Dongeun Lee, Duc Nguyen, Jaesik Choi, Thanh Nguyen-Tang, Toan Tran","submitted_at":"2025-03-03T09:12:14Z","abstract_excerpt":"Recent advancements in large language models (LLMs) based on transformer architectures have sparked significant interest in understanding their inner workings. In this paper, we introduce a novel approach to modeling transformer architectures using highly flexible non-autonomous neural ordinary differential equations (ODEs). Our proposed model parameterizes all weights of attention and feed-forward blocks through neural networks, expressing these weights as functions of a continuous layer index. Through spectral analysis of the model's dynamics, we uncover an increase in eigenvalue magnitude t"},"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":"2503.01329","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-03-03T09:12:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e14aa6f6f10dc5211758226fc05a919d141733e67ec57a7eb7cea71763f59cce","abstract_canon_sha256":"040fa51e2c4dd0a6e6b9cf38c8631c44752f3c1292b962d4b98cba83f27300ba"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:49:51.919538Z","signature_b64":"bgRkSMJo3MxfKNNyvonvXIjXr3361sWUpZ8qG2iWie4LTv7H/1O82UtbVq8vWwDv0Noij/OKaFQJ8BNPdMHIDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"669d3eb2c0a0dadf356b110ed3704ddfb85e0c3f3d1f6296844826b6a4db0de9","last_reissued_at":"2026-07-05T10:49:51.919048Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:49:51.919048Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Neural ODE Transformers: Analyzing Internal Dynamics and Adaptive Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Anh Tong, Cheongwoong Kang, David Hall, Dongeun Lee, Duc Nguyen, Jaesik Choi, Thanh Nguyen-Tang, Toan Tran","submitted_at":"2025-03-03T09:12:14Z","abstract_excerpt":"Recent advancements in large language models (LLMs) based on transformer architectures have sparked significant interest in understanding their inner workings. In this paper, we introduce a novel approach to modeling transformer architectures using highly flexible non-autonomous neural ordinary differential equations (ODEs). Our proposed model parameterizes all weights of attention and feed-forward blocks through neural networks, expressing these weights as functions of a continuous layer index. Through spectral analysis of the model's dynamics, we uncover an increase in eigenvalue magnitude t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.01329","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/2503.01329/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":"2503.01329","created_at":"2026-07-05T10:49:51.919107+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.01329v2","created_at":"2026-07-05T10:49:51.919107+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.01329","created_at":"2026-07-05T10:49:51.919107+00:00"},{"alias_kind":"pith_short_12","alias_value":"M2OT5MWAUDNN","created_at":"2026-07-05T10:49:51.919107+00:00"},{"alias_kind":"pith_short_16","alias_value":"M2OT5MWAUDNN6NLL","created_at":"2026-07-05T10:49:51.919107+00:00"},{"alias_kind":"pith_short_8","alias_value":"M2OT5MWA","created_at":"2026-07-05T10:49:51.919107+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.04154","citing_title":"Robust Filter Attention: Self-Attention as Precision-Weighted State Estimation","ref_index":90,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11007","citing_title":"The Transformer as a Polar State Estimator","ref_index":39,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/M2OT5MWAUDNN6NLLCEHNG4CN36","json":"https://pith.science/pith/M2OT5MWAUDNN6NLLCEHNG4CN36.json","graph_json":"https://pith.science/api/pith-number/M2OT5MWAUDNN6NLLCEHNG4CN36/graph.json","events_json":"https://pith.science/api/pith-number/M2OT5MWAUDNN6NLLCEHNG4CN36/events.json","paper":"https://pith.science/paper/M2OT5MWA"},"agent_actions":{"view_html":"https://pith.science/pith/M2OT5MWAUDNN6NLLCEHNG4CN36","download_json":"https://pith.science/pith/M2OT5MWAUDNN6NLLCEHNG4CN36.json","view_paper":"https://pith.science/paper/M2OT5MWA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.01329&json=true","fetch_graph":"https://pith.science/api/pith-number/M2OT5MWAUDNN6NLLCEHNG4CN36/graph.json","fetch_events":"https://pith.science/api/pith-number/M2OT5MWAUDNN6NLLCEHNG4CN36/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M2OT5MWAUDNN6NLLCEHNG4CN36/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M2OT5MWAUDNN6NLLCEHNG4CN36/action/storage_attestation","attest_author":"https://pith.science/pith/M2OT5MWAUDNN6NLLCEHNG4CN36/action/author_attestation","sign_citation":"https://pith.science/pith/M2OT5MWAUDNN6NLLCEHNG4CN36/action/citation_signature","submit_replication":"https://pith.science/pith/M2OT5MWAUDNN6NLLCEHNG4CN36/action/replication_record"}},"created_at":"2026-07-05T10:49:51.919107+00:00","updated_at":"2026-07-05T10:49:51.919107+00:00"}