{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:65ATZXBWA327K2M3RFEL4JPGUL","short_pith_number":"pith:65ATZXBW","schema_version":"1.0","canonical_sha256":"f7413cdc3606f5f5699b8948be25e6a2f8205064d2e58d1f0e1e3bcfee0a1968","source":{"kind":"arxiv","id":"2607.16848","version":1},"attestation_state":"computed","paper":{"title":"Beyond Memory Leaderboards: Evaluating Scientific Memory as Budgeted Context Restoration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"David Finkelstein, Maksim Sheverev, Sergey Nikolenko","submitted_at":"2026-07-18T15:09:58Z","abstract_excerpt":"Long-term memory is becoming a core component of LLM agents, but most memory benchmarks evaluate conversations or compact summaries, while research agents need to restore evidence from full scientific papers. We introduce two full-text scientific-memory benchmarks, Public AI Memory (PAIM; 81 papers, 66 questions) and Public Transformers (PTr; 252 papers, 98 questions). We evaluate eight memory/retrieval systems, including our own proposed Theoria, plus a no-retrieval baseline. Our results show that memory leaderboards are not interpretable without the full protocol: ingestion granularity, raw-"},"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":"2607.16848","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-18T15:09:58Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7f74de42fec1b4460fcbd587eed006a2617bab1aad4aecc138364d7c6afc36ec","abstract_canon_sha256":"cf25e90ae513e4455aa0d3806531899bb0b8d9b359b894074675304e9f843cd2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-21T01:21:02.059656Z","signature_b64":"+jcFJxhjGCh2cvI110e6yomJTvE053q7220uPeEm/zXaaZsr4A0AoM7H2LVTDauGkXXL98wsHU55RRXa8AjYCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f7413cdc3606f5f5699b8948be25e6a2f8205064d2e58d1f0e1e3bcfee0a1968","last_reissued_at":"2026-07-21T01:21:02.058784Z","signature_status":"signed_v1","first_computed_at":"2026-07-21T01:21:02.058784Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond Memory Leaderboards: Evaluating Scientific Memory as Budgeted Context Restoration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"David Finkelstein, Maksim Sheverev, Sergey Nikolenko","submitted_at":"2026-07-18T15:09:58Z","abstract_excerpt":"Long-term memory is becoming a core component of LLM agents, but most memory benchmarks evaluate conversations or compact summaries, while research agents need to restore evidence from full scientific papers. We introduce two full-text scientific-memory benchmarks, Public AI Memory (PAIM; 81 papers, 66 questions) and Public Transformers (PTr; 252 papers, 98 questions). We evaluate eight memory/retrieval systems, including our own proposed Theoria, plus a no-retrieval baseline. Our results show that memory leaderboards are not interpretable without the full protocol: ingestion granularity, raw-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.16848","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/2607.16848/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":"2607.16848","created_at":"2026-07-21T01:21:02.059228+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.16848v1","created_at":"2026-07-21T01:21:02.059228+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.16848","created_at":"2026-07-21T01:21:02.059228+00:00"},{"alias_kind":"pith_short_12","alias_value":"65ATZXBWA327","created_at":"2026-07-21T01:21:02.059228+00:00"},{"alias_kind":"pith_short_16","alias_value":"65ATZXBWA327K2M3","created_at":"2026-07-21T01:21:02.059228+00:00"},{"alias_kind":"pith_short_8","alias_value":"65ATZXBW","created_at":"2026-07-21T01:21:02.059228+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/65ATZXBWA327K2M3RFEL4JPGUL","json":"https://pith.science/pith/65ATZXBWA327K2M3RFEL4JPGUL.json","graph_json":"https://pith.science/api/pith-number/65ATZXBWA327K2M3RFEL4JPGUL/graph.json","events_json":"https://pith.science/api/pith-number/65ATZXBWA327K2M3RFEL4JPGUL/events.json","paper":"https://pith.science/paper/65ATZXBW"},"agent_actions":{"view_html":"https://pith.science/pith/65ATZXBWA327K2M3RFEL4JPGUL","download_json":"https://pith.science/pith/65ATZXBWA327K2M3RFEL4JPGUL.json","view_paper":"https://pith.science/paper/65ATZXBW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.16848&json=true","fetch_graph":"https://pith.science/api/pith-number/65ATZXBWA327K2M3RFEL4JPGUL/graph.json","fetch_events":"https://pith.science/api/pith-number/65ATZXBWA327K2M3RFEL4JPGUL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/65ATZXBWA327K2M3RFEL4JPGUL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/65ATZXBWA327K2M3RFEL4JPGUL/action/storage_attestation","attest_author":"https://pith.science/pith/65ATZXBWA327K2M3RFEL4JPGUL/action/author_attestation","sign_citation":"https://pith.science/pith/65ATZXBWA327K2M3RFEL4JPGUL/action/citation_signature","submit_replication":"https://pith.science/pith/65ATZXBWA327K2M3RFEL4JPGUL/action/replication_record"}},"created_at":"2026-07-21T01:21:02.059228+00:00","updated_at":"2026-07-21T01:21:02.059228+00:00"}