{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:D5TMFWZYH25S2AEPQMCWOPP4XR","short_pith_number":"pith:D5TMFWZY","schema_version":"1.0","canonical_sha256":"1f66c2db383ebb2d008f8305673dfcbc517fe433b7cd52769127d86000c889e3","source":{"kind":"arxiv","id":"2605.28742","version":1},"attestation_state":"computed","paper":{"title":"CORE: Contrastive Reflection Enables Rapid Improvements in Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Ben Prystawski, Judith E. Fan, Linas Nasvytis, Noah D. Goodman, Satchel Grant, Simon Jerome Han","submitted_at":"2026-05-27T17:01:50Z","abstract_excerpt":"Language models can use verifiable rewards to improve at a wide variety of reasoning tasks. However, both parametric (e.g. RLVR) and non-parametric (e.g. prompt optimization) approaches to doing so typically require hundreds of training samples and thousands of model rollouts, making them expensive in the best case and intractable in the worst. To address this challenge, we introduce Contrastive Reflection (CORE), a non-parametric learning algorithm that compares past reasoning traces to generate insights: short natural-language descriptions of reasoning strategies and constraints that capture"},"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":"2605.28742","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-27T17:01:50Z","cross_cats_sorted":[],"title_canon_sha256":"4444a48f1f803218506328bba5f211f0686526103ab1fe68bb4ff5a8be3b845b","abstract_canon_sha256":"1fa5872438dce17e00a53ec188b11903d23d282af4769079d11a220a82b7c562"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T02:05:01.715710Z","signature_b64":"Jw4E5SashXu1In6Yajl7gUMRM5IxpUGe40LBv4eRTL+aTIllXGBTXXJOYonuWn/0UhRFEr/v/jdgzUV1Oc5+Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1f66c2db383ebb2d008f8305673dfcbc517fe433b7cd52769127d86000c889e3","last_reissued_at":"2026-05-28T02:05:01.715263Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T02:05:01.715263Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CORE: Contrastive Reflection Enables Rapid Improvements in Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Ben Prystawski, Judith E. Fan, Linas Nasvytis, Noah D. Goodman, Satchel Grant, Simon Jerome Han","submitted_at":"2026-05-27T17:01:50Z","abstract_excerpt":"Language models can use verifiable rewards to improve at a wide variety of reasoning tasks. However, both parametric (e.g. RLVR) and non-parametric (e.g. prompt optimization) approaches to doing so typically require hundreds of training samples and thousands of model rollouts, making them expensive in the best case and intractable in the worst. To address this challenge, we introduce Contrastive Reflection (CORE), a non-parametric learning algorithm that compares past reasoning traces to generate insights: short natural-language descriptions of reasoning strategies and constraints that capture"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28742","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/2605.28742/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":"2605.28742","created_at":"2026-05-28T02:05:01.715325+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28742v1","created_at":"2026-05-28T02:05:01.715325+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28742","created_at":"2026-05-28T02:05:01.715325+00:00"},{"alias_kind":"pith_short_12","alias_value":"D5TMFWZYH25S","created_at":"2026-05-28T02:05:01.715325+00:00"},{"alias_kind":"pith_short_16","alias_value":"D5TMFWZYH25S2AEP","created_at":"2026-05-28T02:05:01.715325+00:00"},{"alias_kind":"pith_short_8","alias_value":"D5TMFWZY","created_at":"2026-05-28T02:05:01.715325+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/D5TMFWZYH25S2AEPQMCWOPP4XR","json":"https://pith.science/pith/D5TMFWZYH25S2AEPQMCWOPP4XR.json","graph_json":"https://pith.science/api/pith-number/D5TMFWZYH25S2AEPQMCWOPP4XR/graph.json","events_json":"https://pith.science/api/pith-number/D5TMFWZYH25S2AEPQMCWOPP4XR/events.json","paper":"https://pith.science/paper/D5TMFWZY"},"agent_actions":{"view_html":"https://pith.science/pith/D5TMFWZYH25S2AEPQMCWOPP4XR","download_json":"https://pith.science/pith/D5TMFWZYH25S2AEPQMCWOPP4XR.json","view_paper":"https://pith.science/paper/D5TMFWZY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28742&json=true","fetch_graph":"https://pith.science/api/pith-number/D5TMFWZYH25S2AEPQMCWOPP4XR/graph.json","fetch_events":"https://pith.science/api/pith-number/D5TMFWZYH25S2AEPQMCWOPP4XR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/D5TMFWZYH25S2AEPQMCWOPP4XR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/D5TMFWZYH25S2AEPQMCWOPP4XR/action/storage_attestation","attest_author":"https://pith.science/pith/D5TMFWZYH25S2AEPQMCWOPP4XR/action/author_attestation","sign_citation":"https://pith.science/pith/D5TMFWZYH25S2AEPQMCWOPP4XR/action/citation_signature","submit_replication":"https://pith.science/pith/D5TMFWZYH25S2AEPQMCWOPP4XR/action/replication_record"}},"created_at":"2026-05-28T02:05:01.715325+00:00","updated_at":"2026-05-28T02:05:01.715325+00:00"}