{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:JQNSD25VQTFNQJRJSHZAXYTUEJ","short_pith_number":"pith:JQNSD25V","canonical_record":{"source":{"id":"2504.11972","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-04-16T11:08:46Z","cross_cats_sorted":[],"title_canon_sha256":"495b477d1e1aa5f28d7ed018a570cc1c3dad38450dc1b07c7c83978873752b1f","abstract_canon_sha256":"2656d2a1fc52d5b6c98ca010718ebbcabbeb9e18b30ace7f66b23e3ee1c24639"},"schema_version":"1.0"},"canonical_sha256":"4c1b21ebb584cad8262991f20be2742258e0cda3e89900335ab3a7aa6703bbf4","source":{"kind":"arxiv","id":"2504.11972","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.11972","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"arxiv_version","alias_value":"2504.11972v3","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.11972","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"pith_short_12","alias_value":"JQNSD25VQTFN","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"pith_short_16","alias_value":"JQNSD25VQTFNQJRJ","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"pith_short_8","alias_value":"JQNSD25V","created_at":"2026-06-01T01:02:15Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:JQNSD25VQTFNQJRJSHZAXYTUEJ","target":"record","payload":{"canonical_record":{"source":{"id":"2504.11972","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-04-16T11:08:46Z","cross_cats_sorted":[],"title_canon_sha256":"495b477d1e1aa5f28d7ed018a570cc1c3dad38450dc1b07c7c83978873752b1f","abstract_canon_sha256":"2656d2a1fc52d5b6c98ca010718ebbcabbeb9e18b30ace7f66b23e3ee1c24639"},"schema_version":"1.0"},"canonical_sha256":"4c1b21ebb584cad8262991f20be2742258e0cda3e89900335ab3a7aa6703bbf4","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T01:02:15.883715Z","signature_b64":"yI4uVn9LNXUbJmoK1LeVAeqM7tGzUdH+PnGq93ljWtsHDIviDHycdCOCplIsmFErFowEy4w/3V5EKYtrT9M3Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4c1b21ebb584cad8262991f20be2742258e0cda3e89900335ab3a7aa6703bbf4","last_reissued_at":"2026-06-01T01:02:15.882626Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T01:02:15.882626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2504.11972","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-01T01:02:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"fCwxnTEPz4Fb+tYRRYAx4P0xjUjVdSpTesvNUYAszmTB6hsfO6mOvfU+y0qsII5NCNWJsfa5M2ISPOzx0M7HCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-10T06:23:21.606208Z"},"content_sha256":"7b96c9a1627e0530f49423203ed2a832799af22883be9cde5b5c3ae8f7cbd643","schema_version":"1.0","event_id":"sha256:7b96c9a1627e0530f49423203ed2a832799af22883be9cde5b5c3ae8f7cbd643"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:JQNSD25VQTFNQJRJSHZAXYTUEJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Reassessing Extractive QA Datasets at Scale: LLM-as-a-Judge and In-Depth Analyses","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Akiko Aizawa, Florian Boudin, Jiahao Huang, Xanh Ho","submitted_at":"2025-04-16T11:08:46Z","abstract_excerpt":"Extractive QA tasks are commonly evaluated using Exact Match (EM) and F1-score, but these metrics often fail to reflect true model performance. Recent studies have proposed using large language models (LLMs) as judges (LLM-as-a-judge), yet they often lack comprehensive evaluation across datasets and overlook key factors such as sensitivity to answer types, prompt variations, and self-preference bias. In this work, we conduct a systematic study of LLM-as-a-judge across four extractive QA datasets and various prompt variations, assessing multiple LLM families in both answering and judging roles."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.11972","kind":"arxiv","version":3},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2504.11972/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-01T01:02:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WkfgM1oaoafVw0meuWcTRmTdfDzNxIY7z8odkPEQIgnac5vUrx6n2NLZ9l/4cDdADQtBW1BsQ1zTP0Vih4wpDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-10T06:23:21.607016Z"},"content_sha256":"1f5ca23727c4dbf35f5c8dcb90c4218cab5c92a87c8b430d1676344479a82011","schema_version":"1.0","event_id":"sha256:1f5ca23727c4dbf35f5c8dcb90c4218cab5c92a87c8b430d1676344479a82011"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/JQNSD25VQTFNQJRJSHZAXYTUEJ/bundle.json","state_url":"https://pith.science/pith/JQNSD25VQTFNQJRJSHZAXYTUEJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/JQNSD25VQTFNQJRJSHZAXYTUEJ/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-10T06:23:21Z","links":{"resolver":"https://pith.science/pith/JQNSD25VQTFNQJRJSHZAXYTUEJ","bundle":"https://pith.science/pith/JQNSD25VQTFNQJRJSHZAXYTUEJ/bundle.json","state":"https://pith.science/pith/JQNSD25VQTFNQJRJSHZAXYTUEJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/JQNSD25VQTFNQJRJSHZAXYTUEJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:JQNSD25VQTFNQJRJSHZAXYTUEJ","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":"2656d2a1fc52d5b6c98ca010718ebbcabbeb9e18b30ace7f66b23e3ee1c24639","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-04-16T11:08:46Z","title_canon_sha256":"495b477d1e1aa5f28d7ed018a570cc1c3dad38450dc1b07c7c83978873752b1f"},"schema_version":"1.0","source":{"id":"2504.11972","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.11972","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"arxiv_version","alias_value":"2504.11972v3","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.11972","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"pith_short_12","alias_value":"JQNSD25VQTFN","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"pith_short_16","alias_value":"JQNSD25VQTFNQJRJ","created_at":"2026-06-01T01:02:15Z"},{"alias_kind":"pith_short_8","alias_value":"JQNSD25V","created_at":"2026-06-01T01:02:15Z"}],"graph_snapshots":[{"event_id":"sha256:1f5ca23727c4dbf35f5c8dcb90c4218cab5c92a87c8b430d1676344479a82011","target":"graph","created_at":"2026-06-01T01:02:15Z","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/2504.11972/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Extractive QA tasks are commonly evaluated using Exact Match (EM) and F1-score, but these metrics often fail to reflect true model performance. Recent studies have proposed using large language models (LLMs) as judges (LLM-as-a-judge), yet they often lack comprehensive evaluation across datasets and overlook key factors such as sensitivity to answer types, prompt variations, and self-preference bias. In this work, we conduct a systematic study of LLM-as-a-judge across four extractive QA datasets and various prompt variations, assessing multiple LLM families in both answering and judging roles.","authors_text":"Akiko Aizawa, Florian Boudin, Jiahao Huang, Xanh Ho","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-04-16T11:08:46Z","title":"Reassessing Extractive QA Datasets at Scale: LLM-as-a-Judge and In-Depth Analyses"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.11972","kind":"arxiv","version":3},"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":"sha256:7b96c9a1627e0530f49423203ed2a832799af22883be9cde5b5c3ae8f7cbd643","target":"record","created_at":"2026-06-01T01:02:15Z","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":"2656d2a1fc52d5b6c98ca010718ebbcabbeb9e18b30ace7f66b23e3ee1c24639","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-04-16T11:08:46Z","title_canon_sha256":"495b477d1e1aa5f28d7ed018a570cc1c3dad38450dc1b07c7c83978873752b1f"},"schema_version":"1.0","source":{"id":"2504.11972","kind":"arxiv","version":3}},"canonical_sha256":"4c1b21ebb584cad8262991f20be2742258e0cda3e89900335ab3a7aa6703bbf4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4c1b21ebb584cad8262991f20be2742258e0cda3e89900335ab3a7aa6703bbf4","first_computed_at":"2026-06-01T01:02:15.882626Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-01T01:02:15.882626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"yI4uVn9LNXUbJmoK1LeVAeqM7tGzUdH+PnGq93ljWtsHDIviDHycdCOCplIsmFErFowEy4w/3V5EKYtrT9M3Cg==","signature_status":"signed_v1","signed_at":"2026-06-01T01:02:15.883715Z","signed_message":"canonical_sha256_bytes"},"source_id":"2504.11972","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7b96c9a1627e0530f49423203ed2a832799af22883be9cde5b5c3ae8f7cbd643","sha256:1f5ca23727c4dbf35f5c8dcb90c4218cab5c92a87c8b430d1676344479a82011"],"state_sha256":"5d66e7bbd79bd190575deb927e1973a92c795779060c762ed5e48288696339b2"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"b8yuIi4wrxlIPpPk/dQXEjpwpkxwMPkzTJqMPVxG3puYWSfq+TgtnGrWFsyNyfxt+zTQrj0ZzmEmEoFf/2DGDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-10T06:23:21.611236Z","bundle_sha256":"675bcb6ef9520d1759acaa16a7ebea77c8ebfed941f844b5e5559996b2332421"}}