{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:Y7UCCR4JR3YWV3GDTMSBDMCOY4","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":"3f016c995c01528c728197a357fa50a61da8140b066abab3358cdc968194fbe3","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-09T14:59:14Z","title_canon_sha256":"9399dcdce0d6874fba4faf00d9f3e60691b5c6423ed7cf9aa41727448debcc76"},"schema_version":"1.0","source":{"id":"2606.10956","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.10956","created_at":"2026-06-10T01:10:49Z"},{"alias_kind":"arxiv_version","alias_value":"2606.10956v1","created_at":"2026-06-10T01:10:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.10956","created_at":"2026-06-10T01:10:49Z"},{"alias_kind":"pith_short_12","alias_value":"Y7UCCR4JR3YW","created_at":"2026-06-10T01:10:49Z"},{"alias_kind":"pith_short_16","alias_value":"Y7UCCR4JR3YWV3GD","created_at":"2026-06-10T01:10:49Z"},{"alias_kind":"pith_short_8","alias_value":"Y7UCCR4J","created_at":"2026-06-10T01:10:49Z"}],"graph_snapshots":[{"event_id":"sha256:b304e5053c23cf578f8ce0b9f187627ba49112e8b75d829bbd25d1c899bdbbef","target":"graph","created_at":"2026-06-10T01:10:49Z","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/2606.10956/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The deployment of Large Language Model (LLM) agents for computer automation is accelerating, yet their ability to navigate complex, professional-grade productivity software is largely untested. We argue that Office automation is an ideal environment for benchmarking document-automation capability, as it requires long-horizon planning and reasoning, precise parameter configuration, and multi-application integration. To quantify this capability, we introduce an evaluation based on China's National Computer Rank Examination (NCRE), featuring 200 comprehensive practical-operation tasks across Word","authors_text":"Dongdong Zhang, Furu Wei, Jiayu Ding, Lei Cui, Li Dong, Nan Yang, Shaohan Huang, Tengchao Lv, Wenshan Wu, Xiangyang Zhou, Yilin Jia, Yupan Huang, Yuzhong Zhao","cross_cats":["cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-09T14:59:14Z","title":"Mind the Gap: Can Frontier LLMs Pass a Standardized Office Proficiency Exam?"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.10956","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":"sha256:413975887d693691d49cb765e3b767e3a7392961ed620ff758e51e7dae55aa59","target":"record","created_at":"2026-06-10T01:10:49Z","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":"3f016c995c01528c728197a357fa50a61da8140b066abab3358cdc968194fbe3","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-09T14:59:14Z","title_canon_sha256":"9399dcdce0d6874fba4faf00d9f3e60691b5c6423ed7cf9aa41727448debcc76"},"schema_version":"1.0","source":{"id":"2606.10956","kind":"arxiv","version":1}},"canonical_sha256":"c7e82147898ef16aecc39b2411b04ec725adedb3d96f10b1802ccd7a57d722fc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c7e82147898ef16aecc39b2411b04ec725adedb3d96f10b1802ccd7a57d722fc","first_computed_at":"2026-06-10T01:10:49.989924Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-10T01:10:49.989924Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"1umlXgRR2Jjl/cEAX5CnH/qRRsGzdDWCmRcC406tClDvNDaxN97d/iBPq8N3dVKcMmnBAmGV5YnsfDYZP+sfBg==","signature_status":"signed_v1","signed_at":"2026-06-10T01:10:49.990755Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.10956","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:413975887d693691d49cb765e3b767e3a7392961ed620ff758e51e7dae55aa59","sha256:b304e5053c23cf578f8ce0b9f187627ba49112e8b75d829bbd25d1c899bdbbef"],"state_sha256":"01406571f34785540c82d39bc04c9eb477f1bede130445a621e92746fd2aca85"}