{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:XKENEIE6CKARXURQFSSL3JJ7D6","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":"0b4653d48e5c96edd583665d614e45c03566012a76ae4ea7deb88cae5b6aa212","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2023-05-14T12:31:03Z","title_canon_sha256":"7be1cd5f796e5d9d297d676e7ebdc594c526fe70191b64e46483463933c74bae"},"schema_version":"1.0","source":{"id":"2305.08144","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.08144","created_at":"2026-07-05T08:31:03Z"},{"alias_kind":"arxiv_version","alias_value":"2305.08144v4","created_at":"2026-07-05T08:31:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.08144","created_at":"2026-07-05T08:31:03Z"},{"alias_kind":"pith_short_12","alias_value":"XKENEIE6CKAR","created_at":"2026-07-05T08:31:03Z"},{"alias_kind":"pith_short_16","alias_value":"XKENEIE6CKARXURQ","created_at":"2026-07-05T08:31:03Z"},{"alias_kind":"pith_short_8","alias_value":"XKENEIE6","created_at":"2026-07-05T08:31:03Z"}],"graph_snapshots":[{"event_id":"sha256:03dfebac4199ad96db82989d8637f0a0047e8faea211c9bf854ec7a27cbec513","target":"graph","created_at":"2026-07-05T08:31:03Z","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/2305.08144/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The Graphical User Interface (GUI) is pivotal for human interaction with the digital world, enabling efficient device control and the completion of complex tasks. Recent progress in Large Language Models (LLMs) and Vision Language Models (VLMs) offers the chance to create advanced GUI agents. To ensure their effectiveness, there's a pressing need for qualified benchmarks that provide trustworthy and reproducible evaluations -- a challenge current benchmarks often fail to address. To tackle this issue, we introduce Mobile-Env, a comprehensive toolkit tailored for creating GUI benchmarks in the ","authors_text":"Danyang Zhang, Hongshen Xu, Kai Yu, Lu Chen, Ruisheng Cao, Rui Xie, Situo Zhang, Siyuan Chen, Tianbao Xie, Zhennan Shen, Zihan Zhao","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2023-05-14T12:31:03Z","title":"Mobile-Env: Building Qualified Evaluation Benchmarks for LLM-GUI Interaction"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.08144","kind":"arxiv","version":4},"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":"sha256:88a7f85ef7bd465f6d9d456e876bf468b7339665d5f0ffc01f61e207f147774d","target":"record","created_at":"2026-07-05T08:31:03Z","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":"0b4653d48e5c96edd583665d614e45c03566012a76ae4ea7deb88cae5b6aa212","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2023-05-14T12:31:03Z","title_canon_sha256":"7be1cd5f796e5d9d297d676e7ebdc594c526fe70191b64e46483463933c74bae"},"schema_version":"1.0","source":{"id":"2305.08144","kind":"arxiv","version":4}},"canonical_sha256":"ba88d2209e12811bd2302ca4bda53f1fa19851f6bdb6aba3117f2b618c0d08b6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ba88d2209e12811bd2302ca4bda53f1fa19851f6bdb6aba3117f2b618c0d08b6","first_computed_at":"2026-07-05T08:31:03.606986Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:31:03.606986Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ldKUfKu4z7FG3ToVouTzXqOaZ+ng+WdMT8v5quAPn9k2PfqJBqz2ipgkQDRzs1HSnmNRi81KnV/NYGb2r8rxAA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:31:03.607467Z","signed_message":"canonical_sha256_bytes"},"source_id":"2305.08144","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:88a7f85ef7bd465f6d9d456e876bf468b7339665d5f0ffc01f61e207f147774d","sha256:03dfebac4199ad96db82989d8637f0a0047e8faea211c9bf854ec7a27cbec513"],"state_sha256":"78bf5734af11b86a2f228e113e105a7c1560fc9636bec6df20cca8d5b43f4d4c"}