{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5LVQV4TPBGRXIY6XJM5K2MG3K4","short_pith_number":"pith:5LVQV4TP","schema_version":"1.0","canonical_sha256":"eaeb0af26f09a37463d74b3aad30db5703d9e5411d25980575115502b87e3873","source":{"kind":"arxiv","id":"2406.04291","version":2},"attestation_state":"computed","paper":{"title":"Stratified Prediction-Powered Inference for Hybrid Language Model Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Adam Fisch, Amir Globerson, Bhuwan Dhingra, Joshua Maynez, R. Alex Hofer, William W. Cohen","submitted_at":"2024-06-06T17:37:39Z","abstract_excerpt":"Prediction-powered inference (PPI) is a method that improves statistical estimates based on limited human-labeled data. PPI achieves this by combining small amounts of human-labeled data with larger amounts of data labeled by a reasonably accurate -- but potentially biased -- automatic system, in a way that results in tighter confidence intervals for certain parameters of interest (e.g., the mean performance of a language model). In this paper, we propose a method called Stratified Prediction-Powered Inference (StratPPI), in which we show that the basic PPI estimates can be considerably improv"},"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":"2406.04291","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-06T17:37:39Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"ca9491a3d5af07e6562a71ef3ac3859d54ade1d6a034ce5b6cf4d5745cbacaba","abstract_canon_sha256":"76e6244dd95274e0f3463d91c21131bd0abed6f229ff2faa296ef323fb576ce0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:44:10.018361Z","signature_b64":"+Z6KnwE18oeM4AMpTE2c0cKqBX5GP+M5cZgFSK6TH/9hfE+/MlyjTqcuxRcNcf/hz6GUyZQ2ke++xrX5UkY2Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eaeb0af26f09a37463d74b3aad30db5703d9e5411d25980575115502b87e3873","last_reissued_at":"2026-07-05T09:44:10.017839Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:44:10.017839Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stratified Prediction-Powered Inference for Hybrid Language Model Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Adam Fisch, Amir Globerson, Bhuwan Dhingra, Joshua Maynez, R. Alex Hofer, William W. Cohen","submitted_at":"2024-06-06T17:37:39Z","abstract_excerpt":"Prediction-powered inference (PPI) is a method that improves statistical estimates based on limited human-labeled data. PPI achieves this by combining small amounts of human-labeled data with larger amounts of data labeled by a reasonably accurate -- but potentially biased -- automatic system, in a way that results in tighter confidence intervals for certain parameters of interest (e.g., the mean performance of a language model). In this paper, we propose a method called Stratified Prediction-Powered Inference (StratPPI), in which we show that the basic PPI estimates can be considerably improv"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.04291","kind":"arxiv","version":2},"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/2406.04291/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":"2406.04291","created_at":"2026-07-05T09:44:10.017902+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.04291v2","created_at":"2026-07-05T09:44:10.017902+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.04291","created_at":"2026-07-05T09:44:10.017902+00:00"},{"alias_kind":"pith_short_12","alias_value":"5LVQV4TPBGRX","created_at":"2026-07-05T09:44:10.017902+00:00"},{"alias_kind":"pith_short_16","alias_value":"5LVQV4TPBGRXIY6X","created_at":"2026-07-05T09:44:10.017902+00:00"},{"alias_kind":"pith_short_8","alias_value":"5LVQV4TP","created_at":"2026-07-05T09:44:10.017902+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.09946","citing_title":"Structure from Strategic Interaction & Uncertainty: Risk Sensitive Games for Robust Preference Learning","ref_index":89,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09946","citing_title":"Structure from Strategic Interaction & Uncertainty: Risk Sensitive Games for Robust Preference Learning","ref_index":89,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09918","citing_title":"NaiAD: Initiate Data-Driven Research for LLM Advertising","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5LVQV4TPBGRXIY6XJM5K2MG3K4","json":"https://pith.science/pith/5LVQV4TPBGRXIY6XJM5K2MG3K4.json","graph_json":"https://pith.science/api/pith-number/5LVQV4TPBGRXIY6XJM5K2MG3K4/graph.json","events_json":"https://pith.science/api/pith-number/5LVQV4TPBGRXIY6XJM5K2MG3K4/events.json","paper":"https://pith.science/paper/5LVQV4TP"},"agent_actions":{"view_html":"https://pith.science/pith/5LVQV4TPBGRXIY6XJM5K2MG3K4","download_json":"https://pith.science/pith/5LVQV4TPBGRXIY6XJM5K2MG3K4.json","view_paper":"https://pith.science/paper/5LVQV4TP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.04291&json=true","fetch_graph":"https://pith.science/api/pith-number/5LVQV4TPBGRXIY6XJM5K2MG3K4/graph.json","fetch_events":"https://pith.science/api/pith-number/5LVQV4TPBGRXIY6XJM5K2MG3K4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5LVQV4TPBGRXIY6XJM5K2MG3K4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5LVQV4TPBGRXIY6XJM5K2MG3K4/action/storage_attestation","attest_author":"https://pith.science/pith/5LVQV4TPBGRXIY6XJM5K2MG3K4/action/author_attestation","sign_citation":"https://pith.science/pith/5LVQV4TPBGRXIY6XJM5K2MG3K4/action/citation_signature","submit_replication":"https://pith.science/pith/5LVQV4TPBGRXIY6XJM5K2MG3K4/action/replication_record"}},"created_at":"2026-07-05T09:44:10.017902+00:00","updated_at":"2026-07-05T09:44:10.017902+00:00"}