{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ZLUTRSCPWNGQ5R7VX3SBZPTH6N","short_pith_number":"pith:ZLUTRSCP","schema_version":"1.0","canonical_sha256":"cae938c84fb34d0ec7f5bee41cbe67f352ef38b44b55b0b5866c6fe31752d4e2","source":{"kind":"arxiv","id":"2605.30916","version":1},"attestation_state":"computed","paper":{"title":"Welfare, Improvability, and Variance: A Principal-Agent Approach to Optimal Benchmark Item Aggregation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.GT","econ.TH"],"primary_cat":"cs.LG","authors_text":"Andreas Haupt, Anka Reuel, Justin Hartenstein, Mykel Kochenderfer, Sanmi Koyejo","submitted_at":"2026-05-29T07:01:38Z","abstract_excerpt":"AI benchmarks have well-documented limitations, with prior work examining contamination, saturation, and construct underspecification. Aggregation has received far less attention: benchmarks are typically summarized by uniformly averaging item-level scores, implicitly treating every test item as equally valuable. We model benchmarking as a multitask principal-agent game and show that the welfare loss from a benchmark is determined jointly by three item-level primitives: alignment with normative welfare priorities, marginal improvability, and performance variance. We translate the theory into a"},"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":"2605.30916","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-29T07:01:38Z","cross_cats_sorted":["cs.GT","econ.TH"],"title_canon_sha256":"99ca4d8388a64094050c418da85b3cb494a1fedfb38eee17d2d1eaf21322f439","abstract_canon_sha256":"0fb088cea39875565c708a10d74a3669a7aa8593f48087a894ea415677a29353"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T01:03:25.704168Z","signature_b64":"IMyBzgnUQ5i6FZ3b3O/smEjY2MO1ztkROSOYPTkrAu6jiKIr8G+g2IENx5zyvfTjjFtQIXCOlsQgXznTNU35AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cae938c84fb34d0ec7f5bee41cbe67f352ef38b44b55b0b5866c6fe31752d4e2","last_reissued_at":"2026-06-01T01:03:25.703369Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T01:03:25.703369Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Welfare, Improvability, and Variance: A Principal-Agent Approach to Optimal Benchmark Item Aggregation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.GT","econ.TH"],"primary_cat":"cs.LG","authors_text":"Andreas Haupt, Anka Reuel, Justin Hartenstein, Mykel Kochenderfer, Sanmi Koyejo","submitted_at":"2026-05-29T07:01:38Z","abstract_excerpt":"AI benchmarks have well-documented limitations, with prior work examining contamination, saturation, and construct underspecification. Aggregation has received far less attention: benchmarks are typically summarized by uniformly averaging item-level scores, implicitly treating every test item as equally valuable. We model benchmarking as a multitask principal-agent game and show that the welfare loss from a benchmark is determined jointly by three item-level primitives: alignment with normative welfare priorities, marginal improvability, and performance variance. We translate the theory into a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30916","kind":"arxiv","version":1},"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/2605.30916/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":"2605.30916","created_at":"2026-06-01T01:03:25.703493+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.30916v1","created_at":"2026-06-01T01:03:25.703493+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30916","created_at":"2026-06-01T01:03:25.703493+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZLUTRSCPWNGQ","created_at":"2026-06-01T01:03:25.703493+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZLUTRSCPWNGQ5R7V","created_at":"2026-06-01T01:03:25.703493+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZLUTRSCP","created_at":"2026-06-01T01:03:25.703493+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZLUTRSCPWNGQ5R7VX3SBZPTH6N","json":"https://pith.science/pith/ZLUTRSCPWNGQ5R7VX3SBZPTH6N.json","graph_json":"https://pith.science/api/pith-number/ZLUTRSCPWNGQ5R7VX3SBZPTH6N/graph.json","events_json":"https://pith.science/api/pith-number/ZLUTRSCPWNGQ5R7VX3SBZPTH6N/events.json","paper":"https://pith.science/paper/ZLUTRSCP"},"agent_actions":{"view_html":"https://pith.science/pith/ZLUTRSCPWNGQ5R7VX3SBZPTH6N","download_json":"https://pith.science/pith/ZLUTRSCPWNGQ5R7VX3SBZPTH6N.json","view_paper":"https://pith.science/paper/ZLUTRSCP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.30916&json=true","fetch_graph":"https://pith.science/api/pith-number/ZLUTRSCPWNGQ5R7VX3SBZPTH6N/graph.json","fetch_events":"https://pith.science/api/pith-number/ZLUTRSCPWNGQ5R7VX3SBZPTH6N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZLUTRSCPWNGQ5R7VX3SBZPTH6N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZLUTRSCPWNGQ5R7VX3SBZPTH6N/action/storage_attestation","attest_author":"https://pith.science/pith/ZLUTRSCPWNGQ5R7VX3SBZPTH6N/action/author_attestation","sign_citation":"https://pith.science/pith/ZLUTRSCPWNGQ5R7VX3SBZPTH6N/action/citation_signature","submit_replication":"https://pith.science/pith/ZLUTRSCPWNGQ5R7VX3SBZPTH6N/action/replication_record"}},"created_at":"2026-06-01T01:03:25.703493+00:00","updated_at":"2026-06-01T01:03:25.703493+00:00"}