{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:S3YLKKYDK3XIGCLCPYGO4R6KOJ","short_pith_number":"pith:S3YLKKYD","schema_version":"1.0","canonical_sha256":"96f0b52b0356ee8309627e0cee47ca724246fc5fd983f4b11c8f030efc0f92f2","source":{"kind":"arxiv","id":"2102.01065","version":3},"attestation_state":"computed","paper":{"title":"Do Question Answering Modeling Improvements Hold Across Benchmarks?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Nelson F. Liu, Percy Liang, Robin Jia, Tony Lee","submitted_at":"2021-02-01T18:55:38Z","abstract_excerpt":"Do question answering (QA) modeling improvements (e.g., choice of architecture and training procedure) hold consistently across the diverse landscape of QA benchmarks? To study this question, we introduce the notion of concurrence -- two benchmarks have high concurrence on a set of modeling approaches if they rank the modeling approaches similarly. We measure the concurrence between 32 QA benchmarks on a set of 20 diverse modeling approaches and find that human-constructed benchmarks have high concurrence amongst themselves, even if their passage and question distributions are very different. "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2102.01065","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-02-01T18:55:38Z","cross_cats_sorted":[],"title_canon_sha256":"3337a745cdc751d0183565a4cd638105d3d78deb20922b85ee0f0921c63a504f","abstract_canon_sha256":"592808cbbd2ff1d55116c8145331593a3d23b617f73b187542eb2a7ef3926ade"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:15:40.952731Z","signature_b64":"nmmLBh54xNZq0plyQEvkiAvepKUFLgfYESV0Pj1ous8xjTltDHUJ2sTPiflHfluAuOe5fETpekBuGBP7c5gWDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"96f0b52b0356ee8309627e0cee47ca724246fc5fd983f4b11c8f030efc0f92f2","last_reissued_at":"2026-07-05T06:15:40.952271Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:15:40.952271Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do Question Answering Modeling Improvements Hold Across Benchmarks?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Nelson F. Liu, Percy Liang, Robin Jia, Tony Lee","submitted_at":"2021-02-01T18:55:38Z","abstract_excerpt":"Do question answering (QA) modeling improvements (e.g., choice of architecture and training procedure) hold consistently across the diverse landscape of QA benchmarks? To study this question, we introduce the notion of concurrence -- two benchmarks have high concurrence on a set of modeling approaches if they rank the modeling approaches similarly. We measure the concurrence between 32 QA benchmarks on a set of 20 diverse modeling approaches and find that human-constructed benchmarks have high concurrence amongst themselves, even if their passage and question distributions are very different. "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.01065","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/2102.01065/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2102.01065","created_at":"2026-07-05T06:15:40.952341+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.01065v3","created_at":"2026-07-05T06:15:40.952341+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.01065","created_at":"2026-07-05T06:15:40.952341+00:00"},{"alias_kind":"pith_short_12","alias_value":"S3YLKKYDK3XI","created_at":"2026-07-05T06:15:40.952341+00:00"},{"alias_kind":"pith_short_16","alias_value":"S3YLKKYDK3XIGCLC","created_at":"2026-07-05T06:15:40.952341+00:00"},{"alias_kind":"pith_short_8","alias_value":"S3YLKKYD","created_at":"2026-07-05T06:15:40.952341+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.19689","citing_title":"MIMDE: Exploring the Use of Synthetic vs Human Data for Evaluating Multi-Insight Multi-Document Extraction Tasks","ref_index":25,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/S3YLKKYDK3XIGCLCPYGO4R6KOJ","json":"https://pith.science/pith/S3YLKKYDK3XIGCLCPYGO4R6KOJ.json","graph_json":"https://pith.science/api/pith-number/S3YLKKYDK3XIGCLCPYGO4R6KOJ/graph.json","events_json":"https://pith.science/api/pith-number/S3YLKKYDK3XIGCLCPYGO4R6KOJ/events.json","paper":"https://pith.science/paper/S3YLKKYD"},"agent_actions":{"view_html":"https://pith.science/pith/S3YLKKYDK3XIGCLCPYGO4R6KOJ","download_json":"https://pith.science/pith/S3YLKKYDK3XIGCLCPYGO4R6KOJ.json","view_paper":"https://pith.science/paper/S3YLKKYD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.01065&json=true","fetch_graph":"https://pith.science/api/pith-number/S3YLKKYDK3XIGCLCPYGO4R6KOJ/graph.json","fetch_events":"https://pith.science/api/pith-number/S3YLKKYDK3XIGCLCPYGO4R6KOJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/S3YLKKYDK3XIGCLCPYGO4R6KOJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/S3YLKKYDK3XIGCLCPYGO4R6KOJ/action/storage_attestation","attest_author":"https://pith.science/pith/S3YLKKYDK3XIGCLCPYGO4R6KOJ/action/author_attestation","sign_citation":"https://pith.science/pith/S3YLKKYDK3XIGCLCPYGO4R6KOJ/action/citation_signature","submit_replication":"https://pith.science/pith/S3YLKKYDK3XIGCLCPYGO4R6KOJ/action/replication_record"}},"created_at":"2026-07-05T06:15:40.952341+00:00","updated_at":"2026-07-05T06:15:40.952341+00:00"}