{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:JQKA46UQJTRT24TADJQADNUXFV","short_pith_number":"pith:JQKA46UQ","schema_version":"1.0","canonical_sha256":"4c140e7a904ce33d72601a6001b6972d7940de13e9e58de4dc225a4455a61522","source":{"kind":"arxiv","id":"2602.03224","version":2},"attestation_state":"computed","paper":{"title":"TAME: A Trustworthy Test-Time Evolution of Agent Memory with Systematic Benchmarking","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Huichi Zhou, Jiuan Zhou, Kun Shao, Mingang Chen, Yihang Chen, Yongkang Hu, Yuan Xie, Yu Cheng, Yushuo Zhang, Zhaoxia Yin, Zhizhong Zhang","submitted_at":"2026-02-03T07:52:26Z","abstract_excerpt":"Test-time evolution of agent memory represents a pivotal paradigm for advancing AGI, as it strengthens complex reasoning through experience accumulation without requiring parameter updates. However, even during benign task evolution, agent safety alignment remains vulnerable, a phenomenon known as Agent Memory Misevolution. To evaluate this phenomenon, we construct the Trust-Memevo benchmark and find that agents exhibit an overall decline in trustworthiness across multiple tasks during benign task evolution. To address this issue, we propose TAME, a trust-aware memory evolution framework in wh"},"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":"2602.03224","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-02-03T07:52:26Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"0fb4535c4f022b6e480eb60e12b71db78ffc0172c94294559ee8563388871a6b","abstract_canon_sha256":"3efc790324c2ff79b53517f2d5107c02d838c2a8bfcc0e3290fe67ca47473ecc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-09T01:05:13.842897Z","signature_b64":"q3xNVjaCmNShq33f0TOFdxmvOioU1i4Hz4+4Wuw5QPHD28G4/Cz2fZTbZ3e865g21CF17nnpE02IZ1uoE29+BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4c140e7a904ce33d72601a6001b6972d7940de13e9e58de4dc225a4455a61522","last_reissued_at":"2026-06-09T01:05:13.842337Z","signature_status":"signed_v1","first_computed_at":"2026-06-09T01:05:13.842337Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TAME: A Trustworthy Test-Time Evolution of Agent Memory with Systematic Benchmarking","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Huichi Zhou, Jiuan Zhou, Kun Shao, Mingang Chen, Yihang Chen, Yongkang Hu, Yuan Xie, Yu Cheng, Yushuo Zhang, Zhaoxia Yin, Zhizhong Zhang","submitted_at":"2026-02-03T07:52:26Z","abstract_excerpt":"Test-time evolution of agent memory represents a pivotal paradigm for advancing AGI, as it strengthens complex reasoning through experience accumulation without requiring parameter updates. However, even during benign task evolution, agent safety alignment remains vulnerable, a phenomenon known as Agent Memory Misevolution. To evaluate this phenomenon, we construct the Trust-Memevo benchmark and find that agents exhibit an overall decline in trustworthiness across multiple tasks during benign task evolution. To address this issue, we propose TAME, a trust-aware memory evolution framework in wh"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.03224","kind":"arxiv","version":2},"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/2602.03224/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":"2602.03224","created_at":"2026-06-09T01:05:13.842407+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.03224v2","created_at":"2026-06-09T01:05:13.842407+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.03224","created_at":"2026-06-09T01:05:13.842407+00:00"},{"alias_kind":"pith_short_12","alias_value":"JQKA46UQJTRT","created_at":"2026-06-09T01:05:13.842407+00:00"},{"alias_kind":"pith_short_16","alias_value":"JQKA46UQJTRT24TA","created_at":"2026-06-09T01:05:13.842407+00:00"},{"alias_kind":"pith_short_8","alias_value":"JQKA46UQ","created_at":"2026-06-09T01:05:13.842407+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2605.22842","citing_title":"The Misattribution Gap: When Memory Poisoning Looks Like Model Failure in Agentic AI Systems","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2604.15097","citing_title":"From Procedural Skills to Strategy Genes: Towards Experience-Driven Test-Time Evolution","ref_index":16,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JQKA46UQJTRT24TADJQADNUXFV","json":"https://pith.science/pith/JQKA46UQJTRT24TADJQADNUXFV.json","graph_json":"https://pith.science/api/pith-number/JQKA46UQJTRT24TADJQADNUXFV/graph.json","events_json":"https://pith.science/api/pith-number/JQKA46UQJTRT24TADJQADNUXFV/events.json","paper":"https://pith.science/paper/JQKA46UQ"},"agent_actions":{"view_html":"https://pith.science/pith/JQKA46UQJTRT24TADJQADNUXFV","download_json":"https://pith.science/pith/JQKA46UQJTRT24TADJQADNUXFV.json","view_paper":"https://pith.science/paper/JQKA46UQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.03224&json=true","fetch_graph":"https://pith.science/api/pith-number/JQKA46UQJTRT24TADJQADNUXFV/graph.json","fetch_events":"https://pith.science/api/pith-number/JQKA46UQJTRT24TADJQADNUXFV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JQKA46UQJTRT24TADJQADNUXFV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JQKA46UQJTRT24TADJQADNUXFV/action/storage_attestation","attest_author":"https://pith.science/pith/JQKA46UQJTRT24TADJQADNUXFV/action/author_attestation","sign_citation":"https://pith.science/pith/JQKA46UQJTRT24TADJQADNUXFV/action/citation_signature","submit_replication":"https://pith.science/pith/JQKA46UQJTRT24TADJQADNUXFV/action/replication_record"}},"created_at":"2026-06-09T01:05:13.842407+00:00","updated_at":"2026-06-09T01:05:13.842407+00:00"}