{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:4IY2S5WTZITHS267KCB7Z6ASV7","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"2593ab7f95d28890d3a29f6d64616ebdc13ac951ac1f7c5848db2fed361fb52d","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-08-06T15:14:54Z","title_canon_sha256":"37a85cce745dacbb35384bbafa869dda8647acbc6959e05cd5b771c66ef49e6e"},"schema_version":"1.0","source":{"id":"2608.06144","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.06144","created_at":"2026-08-07T01:40:48Z"},{"alias_kind":"arxiv_version","alias_value":"2608.06144v1","created_at":"2026-08-07T01:40:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.06144","created_at":"2026-08-07T01:40:48Z"},{"alias_kind":"pith_short_12","alias_value":"4IY2S5WTZITH","created_at":"2026-08-07T01:40:48Z"},{"alias_kind":"pith_short_16","alias_value":"4IY2S5WTZITHS267","created_at":"2026-08-07T01:40:48Z"},{"alias_kind":"pith_short_8","alias_value":"4IY2S5WT","created_at":"2026-08-07T01:40:48Z"}],"graph_snapshots":[{"event_id":"sha256:f0579d764b86cf5be3e203dc01f977ceb3110226f0f4f1f6dc712e09ac708300","target":"graph","created_at":"2026-08-07T01:40:48Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2608.06144/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Most agent benchmarks evaluate tasks independently and cannot measure whether experience from one task helps with later tasks. Existing self-evolution benchmarks do not jointly cover professional workflows, open-ended deliverables, and multi-aspect evaluation. We introduce FinEvo-Bench, a longitudinal benchmark with 120 real-case-grounded tasks, 20 business scenes across six financial domains. Institution-provided professional procedures define the required operations and constraints. Eligible institution-provided and publicly documented cases supply the task facts. Each scene contains six rel","authors_text":"2), (2) Qwen DianJin Team, Alibaba Cloud Computing), Bo Deng (1, Chenggang Xie (1), Chi Zhang (2) ((1) Beihang University, Chongyang Tao (1), Feng Chen (2), Kang Zhou (2), Lifan Guo (2), Renzhao Liang (1), Xuanren Chen (1)","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-08-06T15:14:54Z","title":"FinEvo-Bench: A Longitudinal Benchmark for Self-Evolving Agents in Professional Financial Workflows"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.06144","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:3945e3d6cdb5585898587de7e96a4857b4842f82d6d73d536752001397c2b545","target":"record","created_at":"2026-08-07T01:40:48Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"2593ab7f95d28890d3a29f6d64616ebdc13ac951ac1f7c5848db2fed361fb52d","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-08-06T15:14:54Z","title_canon_sha256":"37a85cce745dacbb35384bbafa869dda8647acbc6959e05cd5b771c66ef49e6e"},"schema_version":"1.0","source":{"id":"2608.06144","kind":"arxiv","version":1}},"canonical_sha256":"e231a976d3ca26796bdf5083fcf812aff3719e3f5159980203dacf7da7ca2499","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e231a976d3ca26796bdf5083fcf812aff3719e3f5159980203dacf7da7ca2499","first_computed_at":"2026-08-07T01:40:48.908566Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-08-07T01:40:48.908566Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"/k7LHs5yxo2K30uBDfthE5l1YDdANxyurvefWY3LfwXqF50wvs1Ny3ZYJAhSGSYpAvmjNUUqTwjNULrOq64ZDQ==","signature_status":"signed_v1","signed_at":"2026-08-07T01:40:48.910275Z","signed_message":"canonical_sha256_bytes"},"source_id":"2608.06144","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3945e3d6cdb5585898587de7e96a4857b4842f82d6d73d536752001397c2b545","sha256:f0579d764b86cf5be3e203dc01f977ceb3110226f0f4f1f6dc712e09ac708300"],"state_sha256":"90fb7a524cb98de595dd2aacd039575a265b7270b0cf61653aad16fb3ce713b6"}