{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:G7H3WJ4YR63W2D67DKSQIIB6OT","short_pith_number":"pith:G7H3WJ4Y","schema_version":"1.0","canonical_sha256":"37cfbb27988fb76d0fdf1aa504203e74d62bee4b4d10c5cbf74ff706f9756c07","source":{"kind":"arxiv","id":"2305.07615","version":1},"attestation_state":"computed","paper":{"title":"What are the Desired Characteristics of Calibration Sets? Identifying Correlates on Long Form Scientific Summarization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Anna Ostropolets, Bichlien H Nguyen, Budhaditya Deb, Griffin Adams, Jake Smith, No\\'emie Elhadad, Shufang Xie, Tristan Naumann, Yingce Xia, Yuan-Jyue Chen","submitted_at":"2023-05-12T17:08:47Z","abstract_excerpt":"Summarization models often generate text that is poorly calibrated to quality metrics because they are trained to maximize the likelihood of a single reference (MLE). To address this, recent work has added a calibration step, which exposes a model to its own ranked outputs to improve relevance or, in a separate line of work, contrasts positive and negative sets to improve faithfulness. While effective, much of this work has focused on how to generate and optimize these sets. Less is known about why one setup is more effective than another. In this work, we uncover the underlying characteristic"},"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":"2305.07615","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-12T17:08:47Z","cross_cats_sorted":[],"title_canon_sha256":"d02d2a3ea2d294b7cebed0eb0d0d4822772480fe9fc3976083c3e97fa25f4c35","abstract_canon_sha256":"fb1b2de19134e59fbdbb00391023d511709b9b942c1a491f405458fc09e4d41e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:09:38.258796Z","signature_b64":"TMBKycmgeqniqP9hkPhHEYZ0h/+gl5Q4kHii10Mj6zWaFSrrN2vDrwQSzgJZESAY2nWrcmpjxlgTpyKcke+zBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"37cfbb27988fb76d0fdf1aa504203e74d62bee4b4d10c5cbf74ff706f9756c07","last_reissued_at":"2026-07-05T06:09:38.258406Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:09:38.258406Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"What are the Desired Characteristics of Calibration Sets? Identifying Correlates on Long Form Scientific Summarization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Anna Ostropolets, Bichlien H Nguyen, Budhaditya Deb, Griffin Adams, Jake Smith, No\\'emie Elhadad, Shufang Xie, Tristan Naumann, Yingce Xia, Yuan-Jyue Chen","submitted_at":"2023-05-12T17:08:47Z","abstract_excerpt":"Summarization models often generate text that is poorly calibrated to quality metrics because they are trained to maximize the likelihood of a single reference (MLE). To address this, recent work has added a calibration step, which exposes a model to its own ranked outputs to improve relevance or, in a separate line of work, contrasts positive and negative sets to improve faithfulness. While effective, much of this work has focused on how to generate and optimize these sets. Less is known about why one setup is more effective than another. In this work, we uncover the underlying characteristic"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.07615","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/2305.07615/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":"2305.07615","created_at":"2026-07-05T06:09:38.258467+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.07615v1","created_at":"2026-07-05T06:09:38.258467+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.07615","created_at":"2026-07-05T06:09:38.258467+00:00"},{"alias_kind":"pith_short_12","alias_value":"G7H3WJ4YR63W","created_at":"2026-07-05T06:09:38.258467+00:00"},{"alias_kind":"pith_short_16","alias_value":"G7H3WJ4YR63W2D67","created_at":"2026-07-05T06:09:38.258467+00:00"},{"alias_kind":"pith_short_8","alias_value":"G7H3WJ4Y","created_at":"2026-07-05T06:09:38.258467+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/G7H3WJ4YR63W2D67DKSQIIB6OT","json":"https://pith.science/pith/G7H3WJ4YR63W2D67DKSQIIB6OT.json","graph_json":"https://pith.science/api/pith-number/G7H3WJ4YR63W2D67DKSQIIB6OT/graph.json","events_json":"https://pith.science/api/pith-number/G7H3WJ4YR63W2D67DKSQIIB6OT/events.json","paper":"https://pith.science/paper/G7H3WJ4Y"},"agent_actions":{"view_html":"https://pith.science/pith/G7H3WJ4YR63W2D67DKSQIIB6OT","download_json":"https://pith.science/pith/G7H3WJ4YR63W2D67DKSQIIB6OT.json","view_paper":"https://pith.science/paper/G7H3WJ4Y","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.07615&json=true","fetch_graph":"https://pith.science/api/pith-number/G7H3WJ4YR63W2D67DKSQIIB6OT/graph.json","fetch_events":"https://pith.science/api/pith-number/G7H3WJ4YR63W2D67DKSQIIB6OT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/G7H3WJ4YR63W2D67DKSQIIB6OT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/G7H3WJ4YR63W2D67DKSQIIB6OT/action/storage_attestation","attest_author":"https://pith.science/pith/G7H3WJ4YR63W2D67DKSQIIB6OT/action/author_attestation","sign_citation":"https://pith.science/pith/G7H3WJ4YR63W2D67DKSQIIB6OT/action/citation_signature","submit_replication":"https://pith.science/pith/G7H3WJ4YR63W2D67DKSQIIB6OT/action/replication_record"}},"created_at":"2026-07-05T06:09:38.258467+00:00","updated_at":"2026-07-05T06:09:38.258467+00:00"}