{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:3LZRQ3IJSOTTSOYZEEP5TU7WSE","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":"b5d9f5518e89229c83c3020aaffb60f806c0e6998804d04ccdb28381eee21ab0","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-04T12:04:46Z","title_canon_sha256":"1b1474366ef8bba0de60384768a8d507a4700fbbca02d16f46503ed4fad67378"},"schema_version":"1.0","source":{"id":"2501.02266","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.02266","created_at":"2026-07-05T09:57:05Z"},{"alias_kind":"arxiv_version","alias_value":"2501.02266v1","created_at":"2026-07-05T09:57:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.02266","created_at":"2026-07-05T09:57:05Z"},{"alias_kind":"pith_short_12","alias_value":"3LZRQ3IJSOTT","created_at":"2026-07-05T09:57:05Z"},{"alias_kind":"pith_short_16","alias_value":"3LZRQ3IJSOTTSOYZ","created_at":"2026-07-05T09:57:05Z"},{"alias_kind":"pith_short_8","alias_value":"3LZRQ3IJ","created_at":"2026-07-05T09:57:05Z"}],"graph_snapshots":[{"event_id":"sha256:3ff2b78c94b57c17176c208eb5eb724f921b714bfb7315067376c00f706e5bca","target":"graph","created_at":"2026-07-05T09:57:05Z","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/2501.02266/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"This article introduces the first comprehensive benchmark for the Polish language at this scale: LLMzSz{\\L} (LLMs Behind the School Desk). It is based on a coherent collection of Polish national exams, including both academic and professional tests extracted from the archives of the Polish Central Examination Board. It covers 4 types of exams, coming from 154 domains. Altogether, it consists of almost 19k closed-ended questions. We investigate the performance of open-source multilingual, English, and Polish LLMs to verify LLMs' abilities to transfer knowledge between languages. Also, the corre","authors_text":"Filip Grali\\'nski, Jakub Pokrywka, Krzysztof Jassem, Marek Kubis, Micha{\\l} Ciesi\\'o{\\l}ka, Monika Jab{\\l}o\\'nska, Piotr Jab{\\l}o\\'nski, Ryszard Staruch","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-04T12:04:46Z","title":"LLMzSz{\\L}: a comprehensive LLM benchmark for Polish"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.02266","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":"sha256:16366c1b6af1e2df147cf20eb2a69b4ffad5d39535da228b09d501f82a3f605e","target":"record","created_at":"2026-07-05T09:57:05Z","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":"b5d9f5518e89229c83c3020aaffb60f806c0e6998804d04ccdb28381eee21ab0","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-04T12:04:46Z","title_canon_sha256":"1b1474366ef8bba0de60384768a8d507a4700fbbca02d16f46503ed4fad67378"},"schema_version":"1.0","source":{"id":"2501.02266","kind":"arxiv","version":1}},"canonical_sha256":"daf3186d0993a7393b19211fd9d3f691361d7dc526862d976e0fd26a04be7061","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"daf3186d0993a7393b19211fd9d3f691361d7dc526862d976e0fd26a04be7061","first_computed_at":"2026-07-05T09:57:05.062608Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:57:05.062608Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"LP3ZwH/VUc2atFdpb9Fgp5GW3Ao61fB32mUAeDNHjCglsYRE/Brpq9smxoeNQI+RZ97YlDyOh4UWm6G0ZkpbAw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:57:05.062960Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.02266","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:16366c1b6af1e2df147cf20eb2a69b4ffad5d39535da228b09d501f82a3f605e","sha256:3ff2b78c94b57c17176c208eb5eb724f921b714bfb7315067376c00f706e5bca"],"state_sha256":"5d05b92bd9bb06494463c34c431545bf2aaae4a61c18069b437e3c23abadb70f"}