{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UWIMWRUYJHHPR54CUURJP4HCUH","short_pith_number":"pith:UWIMWRUY","schema_version":"1.0","canonical_sha256":"a590cb469849cef8f782a52297f0e2a1ec7445145c80275727444c6a52b4d5b5","source":{"kind":"arxiv","id":"2605.26667","version":1},"attestation_state":"computed","paper":{"title":"MemFail: Stress-Testing Failure Modes of LLM Memory Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Dawn Song, Ishir Garg, Neel Kolhe, Xuandong Zhao","submitted_at":"2026-05-26T08:03:55Z","abstract_excerpt":"Large language model (LLM) agents increasingly rely on external memory systems to remain consistent across long-horizon interactions, but little empirical work has been done to understand the specific failure modes and design choices that these systems present. Existing benchmarks report aggregate question-answering accuracy and treat memory systems as black boxes, making it impossible to attribute an incorrect answer to a particular failure mode of the system. We introduce MemFail, a diagnostic benchmark that isolates the failure modes of modern LLM memory systems. We begin by formalizing mem"},"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":"2605.26667","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-26T08:03:55Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3098fd5b5df4212e33ea5a2763300cb20a887e3af857328dd3dfd00a22a5292a","abstract_canon_sha256":"181768bc4754e52b9946c2beca90b69b50786a203169fc3314c80fd32fba2eeb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:06:04.804358Z","signature_b64":"cnzTtv3mUXEy46rRIoDMKbaoLFUyITjKFuYimDPb0mhAQhUn3GF7N+qo6ErSQbJXo2fjtaspzXdN3sDESY92AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a590cb469849cef8f782a52297f0e2a1ec7445145c80275727444c6a52b4d5b5","last_reissued_at":"2026-05-27T01:06:04.803586Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:06:04.803586Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MemFail: Stress-Testing Failure Modes of LLM Memory Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Dawn Song, Ishir Garg, Neel Kolhe, Xuandong Zhao","submitted_at":"2026-05-26T08:03:55Z","abstract_excerpt":"Large language model (LLM) agents increasingly rely on external memory systems to remain consistent across long-horizon interactions, but little empirical work has been done to understand the specific failure modes and design choices that these systems present. Existing benchmarks report aggregate question-answering accuracy and treat memory systems as black boxes, making it impossible to attribute an incorrect answer to a particular failure mode of the system. We introduce MemFail, a diagnostic benchmark that isolates the failure modes of modern LLM memory systems. We begin by formalizing mem"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.26667","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/2605.26667/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":"2605.26667","created_at":"2026-05-27T01:06:04.803699+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.26667v1","created_at":"2026-05-27T01:06:04.803699+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.26667","created_at":"2026-05-27T01:06:04.803699+00:00"},{"alias_kind":"pith_short_12","alias_value":"UWIMWRUYJHHP","created_at":"2026-05-27T01:06:04.803699+00:00"},{"alias_kind":"pith_short_16","alias_value":"UWIMWRUYJHHPR54C","created_at":"2026-05-27T01:06:04.803699+00:00"},{"alias_kind":"pith_short_8","alias_value":"UWIMWRUY","created_at":"2026-05-27T01:06:04.803699+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/UWIMWRUYJHHPR54CUURJP4HCUH","json":"https://pith.science/pith/UWIMWRUYJHHPR54CUURJP4HCUH.json","graph_json":"https://pith.science/api/pith-number/UWIMWRUYJHHPR54CUURJP4HCUH/graph.json","events_json":"https://pith.science/api/pith-number/UWIMWRUYJHHPR54CUURJP4HCUH/events.json","paper":"https://pith.science/paper/UWIMWRUY"},"agent_actions":{"view_html":"https://pith.science/pith/UWIMWRUYJHHPR54CUURJP4HCUH","download_json":"https://pith.science/pith/UWIMWRUYJHHPR54CUURJP4HCUH.json","view_paper":"https://pith.science/paper/UWIMWRUY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.26667&json=true","fetch_graph":"https://pith.science/api/pith-number/UWIMWRUYJHHPR54CUURJP4HCUH/graph.json","fetch_events":"https://pith.science/api/pith-number/UWIMWRUYJHHPR54CUURJP4HCUH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UWIMWRUYJHHPR54CUURJP4HCUH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UWIMWRUYJHHPR54CUURJP4HCUH/action/storage_attestation","attest_author":"https://pith.science/pith/UWIMWRUYJHHPR54CUURJP4HCUH/action/author_attestation","sign_citation":"https://pith.science/pith/UWIMWRUYJHHPR54CUURJP4HCUH/action/citation_signature","submit_replication":"https://pith.science/pith/UWIMWRUYJHHPR54CUURJP4HCUH/action/replication_record"}},"created_at":"2026-05-27T01:06:04.803699+00:00","updated_at":"2026-05-27T01:06:04.803699+00:00"}