{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6XPJRQIEORDHNN6VAXANKHTNNL","short_pith_number":"pith:6XPJRQIE","schema_version":"1.0","canonical_sha256":"f5de98c104744676b7d505c0d51e6d6af2025ea041e1a1e89bc64c970060c6f2","source":{"kind":"arxiv","id":"2605.24869","version":1},"attestation_state":"computed","paper":{"title":"Lngram: N-gram Conditional Memory in Latent Space","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Guoyang Xia, Lei Ren, Xiaojie Wang, Yunao Zheng","submitted_at":"2026-05-24T04:59:11Z","abstract_excerpt":"Sequence modeling requires both compositional reasoning and local static knowledge retrieval, yet standard Transformers handle both through dense computation. Engram partially decouples retrieval from the backbone, but its token-based keys remain tied to text tokenization and hash compression. We propose Lngram, a latent-space conditional memory module that learns discrete symbols directly from hidden states and performs N-gram lookup over these symbols. This design removes the dependence on tokenizer IDs and naturally extends to non-text modalities. In our evaluated settings, Lngram outperfor"},"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":"2605.24869","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T04:59:11Z","cross_cats_sorted":[],"title_canon_sha256":"bb6739ffb61fee6e22950173280b2e17bb09cef1ea8af1b8bff05493f925c4d4","abstract_canon_sha256":"dea360e6964dd17bb2ec9b6a5601bf15290fc53c0d0fdd67379410018b42da0b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T01:04:02.579286Z","signature_b64":"yMM+eCl3MLiqGmD2O+p1EBZnCzpx4yME4g8VaHW5Kh6mbcuu/YWXHzyAtLcn+cWS1QDySpTTbgn/4/G8nNCICQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f5de98c104744676b7d505c0d51e6d6af2025ea041e1a1e89bc64c970060c6f2","last_reissued_at":"2026-05-26T01:04:02.578432Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T01:04:02.578432Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Lngram: N-gram Conditional Memory in Latent Space","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Guoyang Xia, Lei Ren, Xiaojie Wang, Yunao Zheng","submitted_at":"2026-05-24T04:59:11Z","abstract_excerpt":"Sequence modeling requires both compositional reasoning and local static knowledge retrieval, yet standard Transformers handle both through dense computation. Engram partially decouples retrieval from the backbone, but its token-based keys remain tied to text tokenization and hash compression. We propose Lngram, a latent-space conditional memory module that learns discrete symbols directly from hidden states and performs N-gram lookup over these symbols. This design removes the dependence on tokenizer IDs and naturally extends to non-text modalities. In our evaluated settings, Lngram outperfor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.24869","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/2605.24869/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":"2605.24869","created_at":"2026-05-26T01:04:02.578579+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.24869v1","created_at":"2026-05-26T01:04:02.578579+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.24869","created_at":"2026-05-26T01:04:02.578579+00:00"},{"alias_kind":"pith_short_12","alias_value":"6XPJRQIEORDH","created_at":"2026-05-26T01:04:02.578579+00:00"},{"alias_kind":"pith_short_16","alias_value":"6XPJRQIEORDHNN6V","created_at":"2026-05-26T01:04:02.578579+00:00"},{"alias_kind":"pith_short_8","alias_value":"6XPJRQIE","created_at":"2026-05-26T01:04:02.578579+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6XPJRQIEORDHNN6VAXANKHTNNL","json":"https://pith.science/pith/6XPJRQIEORDHNN6VAXANKHTNNL.json","graph_json":"https://pith.science/api/pith-number/6XPJRQIEORDHNN6VAXANKHTNNL/graph.json","events_json":"https://pith.science/api/pith-number/6XPJRQIEORDHNN6VAXANKHTNNL/events.json","paper":"https://pith.science/paper/6XPJRQIE"},"agent_actions":{"view_html":"https://pith.science/pith/6XPJRQIEORDHNN6VAXANKHTNNL","download_json":"https://pith.science/pith/6XPJRQIEORDHNN6VAXANKHTNNL.json","view_paper":"https://pith.science/paper/6XPJRQIE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.24869&json=true","fetch_graph":"https://pith.science/api/pith-number/6XPJRQIEORDHNN6VAXANKHTNNL/graph.json","fetch_events":"https://pith.science/api/pith-number/6XPJRQIEORDHNN6VAXANKHTNNL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6XPJRQIEORDHNN6VAXANKHTNNL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6XPJRQIEORDHNN6VAXANKHTNNL/action/storage_attestation","attest_author":"https://pith.science/pith/6XPJRQIEORDHNN6VAXANKHTNNL/action/author_attestation","sign_citation":"https://pith.science/pith/6XPJRQIEORDHNN6VAXANKHTNNL/action/citation_signature","submit_replication":"https://pith.science/pith/6XPJRQIEORDHNN6VAXANKHTNNL/action/replication_record"}},"created_at":"2026-05-26T01:04:02.578579+00:00","updated_at":"2026-05-26T01:04:02.578579+00:00"}