{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:5CFSNSLRQZJCZ4JOHNDZZ6D6RS","short_pith_number":"pith:5CFSNSLR","schema_version":"1.0","canonical_sha256":"e88b26c97186522cf12e3b479cf87e8c87a8b2dfab63b41ca6bc1bf2962eeaf4","source":{"kind":"arxiv","id":"2607.25357","version":1},"attestation_state":"computed","paper":{"title":"Raven: High-Recall Sequence Modeling with Sparse Memory Routing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Albert Gu, Arshia Afzal, Aviv Bick, Eric P. Xing, Volkan Cevher","submitted_at":"2026-07-28T07:04:32Z","abstract_excerpt":"Long-context recall in linear-time sequence models highlights a tradeoff in how they write to memory. State-based linear models, such as state-space models (SSMs) and linear Transformers, write densely, updating the entire state for each newly arrived token, which leads to interference and makes specific past tokens hard to recover. Sliding-window attention (SWA) exhibits the opposite behavior: it writes sparsely by storing explicit token representations, but only within a fixed window, so recall drops once the relevant token is evicted. Interpolating between these models, we introduce Raven, "},"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":"2607.25357","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-28T07:04:32Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7d5f81d01ed66252e3270ad428864002d1e3d9e82d27b16bd75e0fa5954e4258","abstract_canon_sha256":"a512bb865d2a38a9a3c1f63ac5f7ce4a87c98a9b803b4ea761abb405eda7adb4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-29T01:25:11.136946Z","signature_b64":"sY4wDrGfH7WFiDVbbpZpVAx3Je+D6Uz7hUHd/nC8bkQ+x6oM5wHYF/aQIZITeTR2zYqFmTY/FTj1nmypgZR0CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e88b26c97186522cf12e3b479cf87e8c87a8b2dfab63b41ca6bc1bf2962eeaf4","last_reissued_at":"2026-07-29T01:25:11.136057Z","signature_status":"signed_v1","first_computed_at":"2026-07-29T01:25:11.136057Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Raven: High-Recall Sequence Modeling with Sparse Memory Routing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Albert Gu, Arshia Afzal, Aviv Bick, Eric P. Xing, Volkan Cevher","submitted_at":"2026-07-28T07:04:32Z","abstract_excerpt":"Long-context recall in linear-time sequence models highlights a tradeoff in how they write to memory. State-based linear models, such as state-space models (SSMs) and linear Transformers, write densely, updating the entire state for each newly arrived token, which leads to interference and makes specific past tokens hard to recover. Sliding-window attention (SWA) exhibits the opposite behavior: it writes sparsely by storing explicit token representations, but only within a fixed window, so recall drops once the relevant token is evicted. Interpolating between these models, we introduce Raven, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.25357","kind":"arxiv","version":1},"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/2607.25357/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":"2607.25357","created_at":"2026-07-29T01:25:11.136528+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.25357v1","created_at":"2026-07-29T01:25:11.136528+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.25357","created_at":"2026-07-29T01:25:11.136528+00:00"},{"alias_kind":"pith_short_12","alias_value":"5CFSNSLRQZJC","created_at":"2026-07-29T01:25:11.136528+00:00"},{"alias_kind":"pith_short_16","alias_value":"5CFSNSLRQZJCZ4JO","created_at":"2026-07-29T01:25:11.136528+00:00"},{"alias_kind":"pith_short_8","alias_value":"5CFSNSLR","created_at":"2026-07-29T01:25:11.136528+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.12435","citing_title":"MARCH: Scaling Recurrent Memory with Content-Routed State Anchors","ref_index":70,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5CFSNSLRQZJCZ4JOHNDZZ6D6RS","json":"https://pith.science/pith/5CFSNSLRQZJCZ4JOHNDZZ6D6RS.json","graph_json":"https://pith.science/api/pith-number/5CFSNSLRQZJCZ4JOHNDZZ6D6RS/graph.json","events_json":"https://pith.science/api/pith-number/5CFSNSLRQZJCZ4JOHNDZZ6D6RS/events.json","paper":"https://pith.science/paper/5CFSNSLR"},"agent_actions":{"view_html":"https://pith.science/pith/5CFSNSLRQZJCZ4JOHNDZZ6D6RS","download_json":"https://pith.science/pith/5CFSNSLRQZJCZ4JOHNDZZ6D6RS.json","view_paper":"https://pith.science/paper/5CFSNSLR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.25357&json=true","fetch_graph":"https://pith.science/api/pith-number/5CFSNSLRQZJCZ4JOHNDZZ6D6RS/graph.json","fetch_events":"https://pith.science/api/pith-number/5CFSNSLRQZJCZ4JOHNDZZ6D6RS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5CFSNSLRQZJCZ4JOHNDZZ6D6RS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5CFSNSLRQZJCZ4JOHNDZZ6D6RS/action/storage_attestation","attest_author":"https://pith.science/pith/5CFSNSLRQZJCZ4JOHNDZZ6D6RS/action/author_attestation","sign_citation":"https://pith.science/pith/5CFSNSLRQZJCZ4JOHNDZZ6D6RS/action/citation_signature","submit_replication":"https://pith.science/pith/5CFSNSLRQZJCZ4JOHNDZZ6D6RS/action/replication_record"}},"created_at":"2026-07-29T01:25:11.136528+00:00","updated_at":"2026-07-29T01:25:11.136528+00:00"}