{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DGOA5WE7ABF67ZQXXFHG3XZK7U","short_pith_number":"pith:DGOA5WE7","schema_version":"1.0","canonical_sha256":"199c0ed89f004befe617b94e6ddf2afd206a11e4d544f452046374ed69676c84","source":{"kind":"arxiv","id":"2605.24930","version":1},"attestation_state":"computed","paper":{"title":"H$^{2}$MT: Semantic Hierarchy-Aware Hierarchical Memory Transformer","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jason Cong, Maryam Haghifam, Yizhou Sun, Zifan He","submitted_at":"2026-05-24T08:22:05Z","abstract_excerpt":"Transformer-based LLMs achieve strong results on many language tasks; however, long inputs remain challenging because context windows are finite, and prefill latency and memory grow rapidly with prompt length. Flat token-stream processing and chunk-based retrieval can therefore spend substantial computation and context budget on text unrelated to the query. Offline-indexed RAG additionally introduces external storage and index management overhead, and typically appends retrieved evidence as raw text, increasing prefill cost and latency. H^{2}MT makes long-context inference structure-aware: it "},"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":"2605.24930","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T08:22:05Z","cross_cats_sorted":[],"title_canon_sha256":"86eeec844bff45ee5b7c39ba8d1608316f8d456f0c077f9993eed44c181201a8","abstract_canon_sha256":"3b5d0b036c733df1922951fccde8dddf468310f89d86ce26eade3afa65e907cc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T01:04:06.257403Z","signature_b64":"wV3+rO5CU81/+CgrpBf5i9TJfCAhsHZrnmpf0tFMcPPT8wrCnFEq3tnce8t7tyWZWEMtvPL1IY6IUP6fFutZCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"199c0ed89f004befe617b94e6ddf2afd206a11e4d544f452046374ed69676c84","last_reissued_at":"2026-05-26T01:04:06.256742Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T01:04:06.256742Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"H$^{2}$MT: Semantic Hierarchy-Aware Hierarchical Memory Transformer","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jason Cong, Maryam Haghifam, Yizhou Sun, Zifan He","submitted_at":"2026-05-24T08:22:05Z","abstract_excerpt":"Transformer-based LLMs achieve strong results on many language tasks; however, long inputs remain challenging because context windows are finite, and prefill latency and memory grow rapidly with prompt length. Flat token-stream processing and chunk-based retrieval can therefore spend substantial computation and context budget on text unrelated to the query. Offline-indexed RAG additionally introduces external storage and index management overhead, and typically appends retrieved evidence as raw text, increasing prefill cost and latency. H^{2}MT makes long-context inference structure-aware: it "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.24930","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/2605.24930/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":"2605.24930","created_at":"2026-05-26T01:04:06.256823+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.24930v1","created_at":"2026-05-26T01:04:06.256823+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.24930","created_at":"2026-05-26T01:04:06.256823+00:00"},{"alias_kind":"pith_short_12","alias_value":"DGOA5WE7ABF6","created_at":"2026-05-26T01:04:06.256823+00:00"},{"alias_kind":"pith_short_16","alias_value":"DGOA5WE7ABF67ZQX","created_at":"2026-05-26T01:04:06.256823+00:00"},{"alias_kind":"pith_short_8","alias_value":"DGOA5WE7","created_at":"2026-05-26T01:04:06.256823+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/DGOA5WE7ABF67ZQXXFHG3XZK7U","json":"https://pith.science/pith/DGOA5WE7ABF67ZQXXFHG3XZK7U.json","graph_json":"https://pith.science/api/pith-number/DGOA5WE7ABF67ZQXXFHG3XZK7U/graph.json","events_json":"https://pith.science/api/pith-number/DGOA5WE7ABF67ZQXXFHG3XZK7U/events.json","paper":"https://pith.science/paper/DGOA5WE7"},"agent_actions":{"view_html":"https://pith.science/pith/DGOA5WE7ABF67ZQXXFHG3XZK7U","download_json":"https://pith.science/pith/DGOA5WE7ABF67ZQXXFHG3XZK7U.json","view_paper":"https://pith.science/paper/DGOA5WE7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.24930&json=true","fetch_graph":"https://pith.science/api/pith-number/DGOA5WE7ABF67ZQXXFHG3XZK7U/graph.json","fetch_events":"https://pith.science/api/pith-number/DGOA5WE7ABF67ZQXXFHG3XZK7U/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DGOA5WE7ABF67ZQXXFHG3XZK7U/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DGOA5WE7ABF67ZQXXFHG3XZK7U/action/storage_attestation","attest_author":"https://pith.science/pith/DGOA5WE7ABF67ZQXXFHG3XZK7U/action/author_attestation","sign_citation":"https://pith.science/pith/DGOA5WE7ABF67ZQXXFHG3XZK7U/action/citation_signature","submit_replication":"https://pith.science/pith/DGOA5WE7ABF67ZQXXFHG3XZK7U/action/replication_record"}},"created_at":"2026-05-26T01:04:06.256823+00:00","updated_at":"2026-05-26T01:04:06.256823+00:00"}