{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:HT6QQ4TB7OGKVKI3UABOZZ4U73","short_pith_number":"pith:HT6QQ4TB","canonical_record":{"source":{"id":"2403.07008","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-09T02:47:11Z","cross_cats_sorted":["cs.AI","cs.CL","stat.ME"],"title_canon_sha256":"4825a5bfdbb40bf4dd097eeeca06730d89bbabdeed7b64dcc922037f5fc7c59f","abstract_canon_sha256":"cfb29c4c262a64da1fc4447bc450c0e8515de1da77deb8084ab5abebae19d345"},"schema_version":"1.0"},"canonical_sha256":"3cfd087261fb8caaa91ba002ece794fec1c1fd4490c05decc7f17e6b34f67585","source":{"kind":"arxiv","id":"2403.07008","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.07008","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"arxiv_version","alias_value":"2403.07008v3","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.07008","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"pith_short_12","alias_value":"HT6QQ4TB7OGK","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"pith_short_16","alias_value":"HT6QQ4TB7OGKVKI3","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"pith_short_8","alias_value":"HT6QQ4TB","created_at":"2026-06-02T02:04:03Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:HT6QQ4TB7OGKVKI3UABOZZ4U73","target":"record","payload":{"canonical_record":{"source":{"id":"2403.07008","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-09T02:47:11Z","cross_cats_sorted":["cs.AI","cs.CL","stat.ME"],"title_canon_sha256":"4825a5bfdbb40bf4dd097eeeca06730d89bbabdeed7b64dcc922037f5fc7c59f","abstract_canon_sha256":"cfb29c4c262a64da1fc4447bc450c0e8515de1da77deb8084ab5abebae19d345"},"schema_version":"1.0"},"canonical_sha256":"3cfd087261fb8caaa91ba002ece794fec1c1fd4490c05decc7f17e6b34f67585","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:03.965864Z","signature_b64":"xyEHSvVB2uYZqTFA3dBpIPs53pyVMQk9/uUDCyXYWhcePMU2GM1dxe5GGwQEGOSDZyYd75yAuJRGrHF5BznjCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3cfd087261fb8caaa91ba002ece794fec1c1fd4490c05decc7f17e6b34f67585","last_reissued_at":"2026-06-02T02:04:03.965445Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:03.965445Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2403.07008","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":"2026-06-02T02:04:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"T7L13dCsKk7x+4WODnVAHQZBPgt4wPVylLHvw2FDnd3c71ijZ9O1kDmdkdK0IMRyl549pnj6jp+7SEjqvQFsBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T21:40:33.664257Z"},"content_sha256":"9473725617ca36008066e05f2fca9baa731595fb2c0a67071c26f1994642cdb4","schema_version":"1.0","event_id":"sha256:9473725617ca36008066e05f2fca9baa731595fb2c0a67071c26f1994642cdb4"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:HT6QQ4TB7OGKVKI3UABOZZ4U73","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AutoEval Done Right: Using Synthetic Data for Model Evaluation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","stat.ME"],"primary_cat":"cs.LG","authors_text":"Anastasios N. Angelopoulos, Jitendra Malik, Michael I. Jordan, Nir Yosef, Pierre Boyeau","submitted_at":"2024-03-09T02:47:11Z","abstract_excerpt":"The evaluation of machine learning models using human-labeled validation data can be expensive and time-consuming. AI-labeled synthetic data can be used to decrease the number of human annotations required for this purpose in a process called autoevaluation. We suggest efficient and statistically principled algorithms for this purpose that improve sample efficiency while remaining unbiased. These algorithms increase the effective human-labeled sample size by up to 50% on experiments with GPT-4."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.07008","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/2403.07008/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":"2026-06-02T02:04:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rI6BJpOhYInciilW9YZDFGn5a/gD2VvTJGDMPHtPVFwsxoKV0bMNTvrh632CoFo9ud3V58O/D1f6Az2ICIwECw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T21:40:33.664667Z"},"content_sha256":"0a1ccd53584b3b4f754fa74617051d080d0b07ebd9a75add03f25ca49d868e62","schema_version":"1.0","event_id":"sha256:0a1ccd53584b3b4f754fa74617051d080d0b07ebd9a75add03f25ca49d868e62"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/HT6QQ4TB7OGKVKI3UABOZZ4U73/bundle.json","state_url":"https://pith.science/pith/HT6QQ4TB7OGKVKI3UABOZZ4U73/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/HT6QQ4TB7OGKVKI3UABOZZ4U73/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":"2026-06-02T21:40:33Z","links":{"resolver":"https://pith.science/pith/HT6QQ4TB7OGKVKI3UABOZZ4U73","bundle":"https://pith.science/pith/HT6QQ4TB7OGKVKI3UABOZZ4U73/bundle.json","state":"https://pith.science/pith/HT6QQ4TB7OGKVKI3UABOZZ4U73/state.json","well_known_bundle":"https://pith.science/.well-known/pith/HT6QQ4TB7OGKVKI3UABOZZ4U73/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:HT6QQ4TB7OGKVKI3UABOZZ4U73","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":"cfb29c4c262a64da1fc4447bc450c0e8515de1da77deb8084ab5abebae19d345","cross_cats_sorted":["cs.AI","cs.CL","stat.ME"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-09T02:47:11Z","title_canon_sha256":"4825a5bfdbb40bf4dd097eeeca06730d89bbabdeed7b64dcc922037f5fc7c59f"},"schema_version":"1.0","source":{"id":"2403.07008","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.07008","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"arxiv_version","alias_value":"2403.07008v3","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.07008","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"pith_short_12","alias_value":"HT6QQ4TB7OGK","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"pith_short_16","alias_value":"HT6QQ4TB7OGKVKI3","created_at":"2026-06-02T02:04:03Z"},{"alias_kind":"pith_short_8","alias_value":"HT6QQ4TB","created_at":"2026-06-02T02:04:03Z"}],"graph_snapshots":[{"event_id":"sha256:0a1ccd53584b3b4f754fa74617051d080d0b07ebd9a75add03f25ca49d868e62","target":"graph","created_at":"2026-06-02T02:04:03Z","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/2403.07008/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The evaluation of machine learning models using human-labeled validation data can be expensive and time-consuming. AI-labeled synthetic data can be used to decrease the number of human annotations required for this purpose in a process called autoevaluation. We suggest efficient and statistically principled algorithms for this purpose that improve sample efficiency while remaining unbiased. These algorithms increase the effective human-labeled sample size by up to 50% on experiments with GPT-4.","authors_text":"Anastasios N. Angelopoulos, Jitendra Malik, Michael I. Jordan, Nir Yosef, Pierre Boyeau","cross_cats":["cs.AI","cs.CL","stat.ME"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-09T02:47:11Z","title":"AutoEval Done Right: Using Synthetic Data for Model Evaluation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.07008","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":"sha256:9473725617ca36008066e05f2fca9baa731595fb2c0a67071c26f1994642cdb4","target":"record","created_at":"2026-06-02T02:04:03Z","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":"cfb29c4c262a64da1fc4447bc450c0e8515de1da77deb8084ab5abebae19d345","cross_cats_sorted":["cs.AI","cs.CL","stat.ME"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-09T02:47:11Z","title_canon_sha256":"4825a5bfdbb40bf4dd097eeeca06730d89bbabdeed7b64dcc922037f5fc7c59f"},"schema_version":"1.0","source":{"id":"2403.07008","kind":"arxiv","version":3}},"canonical_sha256":"3cfd087261fb8caaa91ba002ece794fec1c1fd4490c05decc7f17e6b34f67585","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3cfd087261fb8caaa91ba002ece794fec1c1fd4490c05decc7f17e6b34f67585","first_computed_at":"2026-06-02T02:04:03.965445Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T02:04:03.965445Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"xyEHSvVB2uYZqTFA3dBpIPs53pyVMQk9/uUDCyXYWhcePMU2GM1dxe5GGwQEGOSDZyYd75yAuJRGrHF5BznjCA==","signature_status":"signed_v1","signed_at":"2026-06-02T02:04:03.965864Z","signed_message":"canonical_sha256_bytes"},"source_id":"2403.07008","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9473725617ca36008066e05f2fca9baa731595fb2c0a67071c26f1994642cdb4","sha256:0a1ccd53584b3b4f754fa74617051d080d0b07ebd9a75add03f25ca49d868e62"],"state_sha256":"e1430ca5551a03c6355bd28a0d157a6f466df14aa7ebab3f1e1f7a93616b3472"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oPnaIfYLoV3uijqFo6U0eXEj7Wsmg5h2xD5ELKSIFvJ4FThrPH9i2fxEN23vVkguc9k4Aeq94rmDNgFif3k1Cg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-02T21:40:33.667277Z","bundle_sha256":"954d979e42b1ab1b4c08162327db7aa347eb4f141a588778d94d2d632852c7d2"}}