{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:WI3SVM3Q4DHDB4NJFBMNMLB3TL","short_pith_number":"pith:WI3SVM3Q","schema_version":"1.0","canonical_sha256":"b2372ab370e0ce30f1a92858d62c3b9adbe8d758e81764ff2677bb229965e9a6","source":{"kind":"arxiv","id":"2501.12352","version":3},"attestation_state":"computed","paper":{"title":"Test-time regression: a unifying framework for designing sequence models with associative memory","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"Emily B. Fox, Jiaxin Shi, Ke Alexander Wang","submitted_at":"2025-01-21T18:32:31Z","abstract_excerpt":"Sequence models lie at the heart of modern deep learning. However, rapid advancements have produced a diversity of seemingly unrelated architectures, such as Transformers and recurrent alternatives. In this paper, we introduce a unifying framework to understand and derive these sequence models, inspired by the empirical importance of associative recall, the capability to retrieve contextually relevant tokens. We formalize associative recall as a two-step process, memorization and retrieval, casting memorization as a regression problem. Layers that combine these two steps perform associative re"},"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":"2501.12352","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-01-21T18:32:31Z","cross_cats_sorted":["cs.AI","cs.NE","stat.ML"],"title_canon_sha256":"50657390a82eee24145a47b4fc97cf968caddb752e247efa680e39fadac27efd","abstract_canon_sha256":"1b2e558e0a36f3773efedd0365b6ddda1e882c7ce43293b1ff1ccfec24a86bd9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:57:32.908949Z","signature_b64":"dw6UnAf0+110a8xerMNx0jgfJNgLggk0Qcyshac/+aqbVQoS9I7ly1rEgG1z6azSwW4FDEG4syfzXn1AMpXUAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b2372ab370e0ce30f1a92858d62c3b9adbe8d758e81764ff2677bb229965e9a6","last_reissued_at":"2026-07-05T10:57:32.908449Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:57:32.908449Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Test-time regression: a unifying framework for designing sequence models with associative memory","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"Emily B. Fox, Jiaxin Shi, Ke Alexander Wang","submitted_at":"2025-01-21T18:32:31Z","abstract_excerpt":"Sequence models lie at the heart of modern deep learning. However, rapid advancements have produced a diversity of seemingly unrelated architectures, such as Transformers and recurrent alternatives. In this paper, we introduce a unifying framework to understand and derive these sequence models, inspired by the empirical importance of associative recall, the capability to retrieve contextually relevant tokens. We formalize associative recall as a two-step process, memorization and retrieval, casting memorization as a regression problem. Layers that combine these two steps perform associative re"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.12352","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/2501.12352/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":"2501.12352","created_at":"2026-07-05T10:57:32.908507+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.12352v3","created_at":"2026-07-05T10:57:32.908507+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.12352","created_at":"2026-07-05T10:57:32.908507+00:00"},{"alias_kind":"pith_short_12","alias_value":"WI3SVM3Q4DHD","created_at":"2026-07-05T10:57:32.908507+00:00"},{"alias_kind":"pith_short_16","alias_value":"WI3SVM3Q4DHDB4NJ","created_at":"2026-07-05T10:57:32.908507+00:00"},{"alias_kind":"pith_short_8","alias_value":"WI3SVM3Q","created_at":"2026-07-05T10:57:32.908507+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":18,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25342","citing_title":"Lifelong In-Context Learning with Transformers Requires Parametric Forms of Attention","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2606.23670","citing_title":"Tapered Language Models","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03825","citing_title":"Dynamic Short Convolutions Improve Transformers","ref_index":175,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25949","citing_title":"Small Models, Strong Priors: Architectural Inductive Bias for Parameter-Efficient Neural PDE Solvers","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2511.21016","citing_title":"Gated KalmaNet: A Fading Memory Layer Through Test-Time Ridge Regression","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2509.22321","citing_title":"Distributed Associative Memory via Online Convex Optimization","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2509.22630","citing_title":"StateX: Enhancing RNN Recall via Post-training State Expansion","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2509.26645","citing_title":"TTT3R: 3D Reconstruction as Test-Time Training","ref_index":88,"is_internal_anchor":false},{"citing_arxiv_id":"2601.22766","citing_title":"Sparse Attention as Compact Kernel Regression","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2602.21204","citing_title":"Test-Time Training with KV Binding Is Secretly Linear Attention","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2603.04385","citing_title":"ZipMap: Linear-Time Stateful 3D Reconstruction via Test-Time Training","ref_index":70,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13473","citing_title":"OSDN: Improving Delta Rule with Provable Online Preconditioning in Linear Attention","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08587","citing_title":"Kaczmarz Linear Attention","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05838","citing_title":"MDN: Parallelizing Stepwise Momentum for Delta Linear Attention","ref_index":84,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08542","citing_title":"Scal3R: Scalable Test-Time Training for Large-Scale 3D Reconstruction","ref_index":79,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07350","citing_title":"Fast Spatial Memory with Elastic Test-Time Training","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06169","citing_title":"In-Place Test-Time Training","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21100","citing_title":"Preconditioned DeltaNet: Curvature-aware Sequence Modeling for Linear Recurrences","ref_index":56,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WI3SVM3Q4DHDB4NJFBMNMLB3TL","json":"https://pith.science/pith/WI3SVM3Q4DHDB4NJFBMNMLB3TL.json","graph_json":"https://pith.science/api/pith-number/WI3SVM3Q4DHDB4NJFBMNMLB3TL/graph.json","events_json":"https://pith.science/api/pith-number/WI3SVM3Q4DHDB4NJFBMNMLB3TL/events.json","paper":"https://pith.science/paper/WI3SVM3Q"},"agent_actions":{"view_html":"https://pith.science/pith/WI3SVM3Q4DHDB4NJFBMNMLB3TL","download_json":"https://pith.science/pith/WI3SVM3Q4DHDB4NJFBMNMLB3TL.json","view_paper":"https://pith.science/paper/WI3SVM3Q","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.12352&json=true","fetch_graph":"https://pith.science/api/pith-number/WI3SVM3Q4DHDB4NJFBMNMLB3TL/graph.json","fetch_events":"https://pith.science/api/pith-number/WI3SVM3Q4DHDB4NJFBMNMLB3TL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WI3SVM3Q4DHDB4NJFBMNMLB3TL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WI3SVM3Q4DHDB4NJFBMNMLB3TL/action/storage_attestation","attest_author":"https://pith.science/pith/WI3SVM3Q4DHDB4NJFBMNMLB3TL/action/author_attestation","sign_citation":"https://pith.science/pith/WI3SVM3Q4DHDB4NJFBMNMLB3TL/action/citation_signature","submit_replication":"https://pith.science/pith/WI3SVM3Q4DHDB4NJFBMNMLB3TL/action/replication_record"}},"created_at":"2026-07-05T10:57:32.908507+00:00","updated_at":"2026-07-05T10:57:32.908507+00:00"}