{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IQLY2LWVYWUGF5WZ3E5Y6OUPMF","short_pith_number":"pith:IQLY2LWV","schema_version":"1.0","canonical_sha256":"44178d2ed5c5a862f6d9d93b8f3a8f615ebf0b90c5a08be88b8b4576c37c375d","source":{"kind":"arxiv","id":"2404.08819","version":3},"attestation_state":"computed","paper":{"title":"The Illusion of State in State-Space Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.CL","cs.FL"],"primary_cat":"cs.LG","authors_text":"Ashish Sabharwal, Jackson Petty, William Merrill","submitted_at":"2024-04-12T21:30:06Z","abstract_excerpt":"State-space models (SSMs) have emerged as a potential alternative architecture for building large language models (LLMs) compared to the previously ubiquitous transformer architecture. One theoretical weakness of transformers is that they cannot express certain kinds of sequential computation and state tracking (Merrill & Sabharwal, 2023), which SSMs are explicitly designed to address via their close architectural similarity to recurrent neural networks (RNNs). But do SSMs truly have an advantage (over transformers) in expressive power for state tracking? Surprisingly, the answer is no. Our an"},"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":"2404.08819","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-12T21:30:06Z","cross_cats_sorted":["cs.CC","cs.CL","cs.FL"],"title_canon_sha256":"a0b7813a44f32118476c1b03a0967a73d3a5fc8683b5abbaf56fb7e9bd76166d","abstract_canon_sha256":"58cfb138c1a4d774d020e96c15fab91681e21f3a24337e336ba7843afb46e777"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:25:05.904254Z","signature_b64":"GJefCsjtG+K0bCW7THEA5XzJS4xi6+tuk3vWWIzKZ1oajd+4lrPbWOppmEOogkDeyHBhbUv1fAkVjEfQp1/gAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"44178d2ed5c5a862f6d9d93b8f3a8f615ebf0b90c5a08be88b8b4576c37c375d","last_reissued_at":"2026-07-05T10:25:05.903621Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:25:05.903621Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Illusion of State in State-Space Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.CL","cs.FL"],"primary_cat":"cs.LG","authors_text":"Ashish Sabharwal, Jackson Petty, William Merrill","submitted_at":"2024-04-12T21:30:06Z","abstract_excerpt":"State-space models (SSMs) have emerged as a potential alternative architecture for building large language models (LLMs) compared to the previously ubiquitous transformer architecture. One theoretical weakness of transformers is that they cannot express certain kinds of sequential computation and state tracking (Merrill & Sabharwal, 2023), which SSMs are explicitly designed to address via their close architectural similarity to recurrent neural networks (RNNs). But do SSMs truly have an advantage (over transformers) in expressive power for state tracking? Surprisingly, the answer is no. Our an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.08819","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/2404.08819/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":"2404.08819","created_at":"2026-07-05T10:25:05.903692+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.08819v3","created_at":"2026-07-05T10:25:05.903692+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.08819","created_at":"2026-07-05T10:25:05.903692+00:00"},{"alias_kind":"pith_short_12","alias_value":"IQLY2LWVYWUG","created_at":"2026-07-05T10:25:05.903692+00:00"},{"alias_kind":"pith_short_16","alias_value":"IQLY2LWVYWUGF5WZ","created_at":"2026-07-05T10:25:05.903692+00:00"},{"alias_kind":"pith_short_8","alias_value":"IQLY2LWV","created_at":"2026-07-05T10:25:05.903692+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25777","citing_title":"Space-Efficient Language Generation in the Limit","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21884","citing_title":"A Verifiable Search Is Not a Learnable Chain-of-Thought","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09917","citing_title":"SPDM: Geometry-Modulated State Space Modeling with Manifold Constraints for Time Series Forecasting","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07254","citing_title":"A Held-Out Transition-Pair Falsifier for Long-Horizon Non-Abelian State Tracking","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06479","citing_title":"Pretraining Recurrent Networks without Recurrence","ref_index":85,"is_internal_anchor":false},{"citing_arxiv_id":"2511.05963","citing_title":"Next-Latent Prediction Transformers Learn Compact World Models","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2408.01129","citing_title":"A Survey of Mamba","ref_index":137,"is_internal_anchor":false},{"citing_arxiv_id":"2507.12549","citing_title":"The Serial Scaling Hypothesis","ref_index":74,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16325","citing_title":"UniMamba: A Unified Spatial-Temporal Modeling Framework with State-Space and Attention Integration","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2510.26692","citing_title":"Kimi Linear: An Expressive, Efficient Attention Architecture","ref_index":65,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06946","citing_title":"Adaptive Memory Decay for Log-Linear Attention","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05923","citing_title":"The UNDO Flip-Flop: A Controlled Probe for Reversible Semantic State Management in State Space Model","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21100","citing_title":"Preconditioned DeltaNet: Curvature-aware Sequence Modeling for Linear Recurrences","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02442","citing_title":"Measuring AI Reasoning: A Guide for Researchers","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IQLY2LWVYWUGF5WZ3E5Y6OUPMF","json":"https://pith.science/pith/IQLY2LWVYWUGF5WZ3E5Y6OUPMF.json","graph_json":"https://pith.science/api/pith-number/IQLY2LWVYWUGF5WZ3E5Y6OUPMF/graph.json","events_json":"https://pith.science/api/pith-number/IQLY2LWVYWUGF5WZ3E5Y6OUPMF/events.json","paper":"https://pith.science/paper/IQLY2LWV"},"agent_actions":{"view_html":"https://pith.science/pith/IQLY2LWVYWUGF5WZ3E5Y6OUPMF","download_json":"https://pith.science/pith/IQLY2LWVYWUGF5WZ3E5Y6OUPMF.json","view_paper":"https://pith.science/paper/IQLY2LWV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.08819&json=true","fetch_graph":"https://pith.science/api/pith-number/IQLY2LWVYWUGF5WZ3E5Y6OUPMF/graph.json","fetch_events":"https://pith.science/api/pith-number/IQLY2LWVYWUGF5WZ3E5Y6OUPMF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IQLY2LWVYWUGF5WZ3E5Y6OUPMF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IQLY2LWVYWUGF5WZ3E5Y6OUPMF/action/storage_attestation","attest_author":"https://pith.science/pith/IQLY2LWVYWUGF5WZ3E5Y6OUPMF/action/author_attestation","sign_citation":"https://pith.science/pith/IQLY2LWVYWUGF5WZ3E5Y6OUPMF/action/citation_signature","submit_replication":"https://pith.science/pith/IQLY2LWVYWUGF5WZ3E5Y6OUPMF/action/replication_record"}},"created_at":"2026-07-05T10:25:05.903692+00:00","updated_at":"2026-07-05T10:25:05.903692+00:00"}