{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:IQU7FBGGJMWQPK5DPAQU3MSNAL","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":"1987799e4b187221d1fa63557be50ba0eda29cbc5e2559b28697000ee58e0054","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-12-31T07:20:32Z","title_canon_sha256":"d888e53ead8d12a21367536cca1e4543a3c902afd21abd17f760bb53a6f94bf4"},"schema_version":"1.0","source":{"id":"2501.00316","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.00316","created_at":"2026-07-05T11:16:53Z"},{"alias_kind":"arxiv_version","alias_value":"2501.00316v2","created_at":"2026-07-05T11:16:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.00316","created_at":"2026-07-05T11:16:53Z"},{"alias_kind":"pith_short_12","alias_value":"IQU7FBGGJMWQ","created_at":"2026-07-05T11:16:53Z"},{"alias_kind":"pith_short_16","alias_value":"IQU7FBGGJMWQPK5D","created_at":"2026-07-05T11:16:53Z"},{"alias_kind":"pith_short_8","alias_value":"IQU7FBGG","created_at":"2026-07-05T11:16:53Z"}],"graph_snapshots":[{"event_id":"sha256:674f6617842c3266e4432a98e909545114fe162e51ab2e1fc812239b4ce397fd","target":"graph","created_at":"2026-07-05T11:16:53Z","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/2501.00316/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advancements in foundation models have improved autonomous tool usage and reasoning, but their capabilities in map-based reasoning remain underexplored. To address this, we introduce MapEval, a benchmark designed to assess foundation models across three distinct tasks - textual, API-based, and visual reasoning - through 700 multiple-choice questions spanning 180 cities and 54 countries, covering spatial relationships, navigation, travel planning, and real-world map interactions. Unlike prior benchmarks that focus on simple location queries, MapEval requires models to handle long-context","authors_text":"Mahir Labib Dihan, Md Almash Alam, Md Hasebul Hasan, Md Rizwan Parvez, MD Tanvir Hassan, Md Tanvir Parvez, Mohammed Eunus Ali, Muhammad Aamir Cheema","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-12-31T07:20:32Z","title":"MapEval: A Map-Based Evaluation of Geo-Spatial Reasoning in Foundation Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.00316","kind":"arxiv","version":2},"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":"sha256:23a9512d74a22d30ea590dbca93163954bfd4c1376b9468bbdb594ee0f4f2ab1","target":"record","created_at":"2026-07-05T11:16:53Z","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":"1987799e4b187221d1fa63557be50ba0eda29cbc5e2559b28697000ee58e0054","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-12-31T07:20:32Z","title_canon_sha256":"d888e53ead8d12a21367536cca1e4543a3c902afd21abd17f760bb53a6f94bf4"},"schema_version":"1.0","source":{"id":"2501.00316","kind":"arxiv","version":2}},"canonical_sha256":"4429f284c64b2d07aba378214db24d02dc493d200ebd4b7aa2e000dc31f873f3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4429f284c64b2d07aba378214db24d02dc493d200ebd4b7aa2e000dc31f873f3","first_computed_at":"2026-07-05T11:16:53.777447Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:16:53.777447Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"LTwQ0sVlzXOmwJf2X54pHL+7XBRKiBBSBcLH0X3bMjVReXg4SHJlpjcSihJLw0i/oHZR6ROuQ+rhnguRciNoDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:16:53.777943Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.00316","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:23a9512d74a22d30ea590dbca93163954bfd4c1376b9468bbdb594ee0f4f2ab1","sha256:674f6617842c3266e4432a98e909545114fe162e51ab2e1fc812239b4ce397fd"],"state_sha256":"6a8825c723465eb29df964cb74cd1d17361cc09b4a21c94c6fa31d6aea52a66f"}