{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BTGJYTPOHLPNWS4YDFGE3HLOCX","short_pith_number":"pith:BTGJYTPO","schema_version":"1.0","canonical_sha256":"0ccc9c4dee3adedb4b98194c4d9d6e15d0b109fe5a146ae1573a06be9c29ff9b","source":{"kind":"arxiv","id":"2607.03528","version":1},"attestation_state":"computed","paper":{"title":"Aligning Language Models with Selective Prediction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Aryan Deshwal, Gaoxiang Luo, Ju Sun, Sinian Zhang, Yifan Wu","submitted_at":"2026-07-03T17:59:02Z","abstract_excerpt":"Large language models (LLMs) are increasingly deployed as critical decision-making components in high-stakes real-world AI systems, rendering LLM reliability a foremost practical concern. In this paper, we focus on enhancing LLM reliability through selective prediction (SP), a strategy that allows an LLM to only predict for inputs where it is likely to be correct (i.e., coverage) and hence reduce the error rate (i.e., risk) on that portion of inputs -- flagging the remaining inputs for future human discretion. In other words, SP improves LLM reliability by balancing the risk-coverage trade-off"},"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":"2607.03528","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-03T17:59:02Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"c80ca5642ab6b95856982be93528e43b45590b551efb5b549789ceaac4ab8f87","abstract_canon_sha256":"e6e7814e59a15f20b085d01b724cfed4cceaa347e5565dff04c3bf85e7592954"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:17:51.924158Z","signature_b64":"CvBoiMsAQsC1YI7MqOImCuYrdPQOD3OPTmlphwVNThtFwWFqub6jjhi0j2xRoEGr9ERTT7bdXfLnBcAW9ItkDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0ccc9c4dee3adedb4b98194c4d9d6e15d0b109fe5a146ae1573a06be9c29ff9b","last_reissued_at":"2026-07-07T02:17:51.923243Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:17:51.923243Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Aligning Language Models with Selective Prediction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Aryan Deshwal, Gaoxiang Luo, Ju Sun, Sinian Zhang, Yifan Wu","submitted_at":"2026-07-03T17:59:02Z","abstract_excerpt":"Large language models (LLMs) are increasingly deployed as critical decision-making components in high-stakes real-world AI systems, rendering LLM reliability a foremost practical concern. In this paper, we focus on enhancing LLM reliability through selective prediction (SP), a strategy that allows an LLM to only predict for inputs where it is likely to be correct (i.e., coverage) and hence reduce the error rate (i.e., risk) on that portion of inputs -- flagging the remaining inputs for future human discretion. In other words, SP improves LLM reliability by balancing the risk-coverage trade-off"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.03528","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/2607.03528/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":"2607.03528","created_at":"2026-07-07T02:17:51.923407+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.03528v1","created_at":"2026-07-07T02:17:51.923407+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.03528","created_at":"2026-07-07T02:17:51.923407+00:00"},{"alias_kind":"pith_short_12","alias_value":"BTGJYTPOHLPN","created_at":"2026-07-07T02:17:51.923407+00:00"},{"alias_kind":"pith_short_16","alias_value":"BTGJYTPOHLPNWS4Y","created_at":"2026-07-07T02:17:51.923407+00:00"},{"alias_kind":"pith_short_8","alias_value":"BTGJYTPO","created_at":"2026-07-07T02:17:51.923407+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/BTGJYTPOHLPNWS4YDFGE3HLOCX","json":"https://pith.science/pith/BTGJYTPOHLPNWS4YDFGE3HLOCX.json","graph_json":"https://pith.science/api/pith-number/BTGJYTPOHLPNWS4YDFGE3HLOCX/graph.json","events_json":"https://pith.science/api/pith-number/BTGJYTPOHLPNWS4YDFGE3HLOCX/events.json","paper":"https://pith.science/paper/BTGJYTPO"},"agent_actions":{"view_html":"https://pith.science/pith/BTGJYTPOHLPNWS4YDFGE3HLOCX","download_json":"https://pith.science/pith/BTGJYTPOHLPNWS4YDFGE3HLOCX.json","view_paper":"https://pith.science/paper/BTGJYTPO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.03528&json=true","fetch_graph":"https://pith.science/api/pith-number/BTGJYTPOHLPNWS4YDFGE3HLOCX/graph.json","fetch_events":"https://pith.science/api/pith-number/BTGJYTPOHLPNWS4YDFGE3HLOCX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BTGJYTPOHLPNWS4YDFGE3HLOCX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BTGJYTPOHLPNWS4YDFGE3HLOCX/action/storage_attestation","attest_author":"https://pith.science/pith/BTGJYTPOHLPNWS4YDFGE3HLOCX/action/author_attestation","sign_citation":"https://pith.science/pith/BTGJYTPOHLPNWS4YDFGE3HLOCX/action/citation_signature","submit_replication":"https://pith.science/pith/BTGJYTPOHLPNWS4YDFGE3HLOCX/action/replication_record"}},"created_at":"2026-07-07T02:17:51.923407+00:00","updated_at":"2026-07-07T02:17:51.923407+00:00"}