{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IM6GIFIZ3MXOJOMER27CZ7T3WS","short_pith_number":"pith:IM6GIFIZ","schema_version":"1.0","canonical_sha256":"433c641519db2ee4b9848ebe2cfe7bb49be8a00f1f5e47564388aaaad8e9303d","source":{"kind":"arxiv","id":"2406.00209","version":3},"attestation_state":"computed","paper":{"title":"Mamba State-Space Models Are Lyapunov-Stable Learners","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"John T. Halloran, Manbir Gulati, Paul F. Roysdon","submitted_at":"2024-05-31T21:46:23Z","abstract_excerpt":"Mamba state-space models (SSMs) have recently outperformed state-of-the-art (SOTA) Transformer large language models (LLMs) in various tasks and been widely adapted. However, a major concern for stable learning in recurrent-based deep models (such as SSMs) is the sensitivity of their recurrent dynamics. Despite widespread adaptation, the sensitivity of Mamba's recurrent dynamics under common fine-tuning methods-e.g., mixed-precision fine-tuning (MPFT) and parameter-efficient fine-tuning (PEFT)-remains unexplored. Empirically, we show that Mamba LLMs are extremely stable to changes introduced b"},"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":"2406.00209","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-31T21:46:23Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"cfd20b3efa33f1f1bbe3ecfe35620f3dfb5d5d56a19e8a18597a2933bd2b149a","abstract_canon_sha256":"b50bf75e2065b705ab3c1b284a4e37ab200bb3d088f599e889b44dd968540cd7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:01:18.434117Z","signature_b64":"sizKAVQBVNoDun8ZEF1IURYNjHP13tXRwNwD3jwAdZ4WzkgpuTz5qZvxO+4qp+rIsQ8mQjccq5b6QyeG0XnvDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"433c641519db2ee4b9848ebe2cfe7bb49be8a00f1f5e47564388aaaad8e9303d","last_reissued_at":"2026-07-05T12:01:18.433624Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:01:18.433624Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mamba State-Space Models Are Lyapunov-Stable Learners","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"John T. Halloran, Manbir Gulati, Paul F. Roysdon","submitted_at":"2024-05-31T21:46:23Z","abstract_excerpt":"Mamba state-space models (SSMs) have recently outperformed state-of-the-art (SOTA) Transformer large language models (LLMs) in various tasks and been widely adapted. However, a major concern for stable learning in recurrent-based deep models (such as SSMs) is the sensitivity of their recurrent dynamics. Despite widespread adaptation, the sensitivity of Mamba's recurrent dynamics under common fine-tuning methods-e.g., mixed-precision fine-tuning (MPFT) and parameter-efficient fine-tuning (PEFT)-remains unexplored. Empirically, we show that Mamba LLMs are extremely stable to changes introduced b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.00209","kind":"arxiv","version":3},"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/2406.00209/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":"2406.00209","created_at":"2026-07-05T12:01:18.433683+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.00209v3","created_at":"2026-07-05T12:01:18.433683+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.00209","created_at":"2026-07-05T12:01:18.433683+00:00"},{"alias_kind":"pith_short_12","alias_value":"IM6GIFIZ3MXO","created_at":"2026-07-05T12:01:18.433683+00:00"},{"alias_kind":"pith_short_16","alias_value":"IM6GIFIZ3MXOJOME","created_at":"2026-07-05T12:01:18.433683+00:00"},{"alias_kind":"pith_short_8","alias_value":"IM6GIFIZ","created_at":"2026-07-05T12:01:18.433683+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.04488","citing_title":"Energy- and Memory-Efficient PEFT Methods for Personalized On-Device SLMs on Consumer GPUs","ref_index":42,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IM6GIFIZ3MXOJOMER27CZ7T3WS","json":"https://pith.science/pith/IM6GIFIZ3MXOJOMER27CZ7T3WS.json","graph_json":"https://pith.science/api/pith-number/IM6GIFIZ3MXOJOMER27CZ7T3WS/graph.json","events_json":"https://pith.science/api/pith-number/IM6GIFIZ3MXOJOMER27CZ7T3WS/events.json","paper":"https://pith.science/paper/IM6GIFIZ"},"agent_actions":{"view_html":"https://pith.science/pith/IM6GIFIZ3MXOJOMER27CZ7T3WS","download_json":"https://pith.science/pith/IM6GIFIZ3MXOJOMER27CZ7T3WS.json","view_paper":"https://pith.science/paper/IM6GIFIZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.00209&json=true","fetch_graph":"https://pith.science/api/pith-number/IM6GIFIZ3MXOJOMER27CZ7T3WS/graph.json","fetch_events":"https://pith.science/api/pith-number/IM6GIFIZ3MXOJOMER27CZ7T3WS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IM6GIFIZ3MXOJOMER27CZ7T3WS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IM6GIFIZ3MXOJOMER27CZ7T3WS/action/storage_attestation","attest_author":"https://pith.science/pith/IM6GIFIZ3MXOJOMER27CZ7T3WS/action/author_attestation","sign_citation":"https://pith.science/pith/IM6GIFIZ3MXOJOMER27CZ7T3WS/action/citation_signature","submit_replication":"https://pith.science/pith/IM6GIFIZ3MXOJOMER27CZ7T3WS/action/replication_record"}},"created_at":"2026-07-05T12:01:18.433683+00:00","updated_at":"2026-07-05T12:01:18.433683+00:00"}