{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4LTYRC5JUDHDMSGHSXWTI4BDZC","short_pith_number":"pith:4LTYRC5J","schema_version":"1.0","canonical_sha256":"e2e7888ba9a0ce3648c795ed347023c88985f8e7ef51fcd84fb7d693101b108f","source":{"kind":"arxiv","id":"2605.19665","version":1},"attestation_state":"computed","paper":{"title":"CriterAlign: Criterion-Centric Rationale Alignment for Code Preference Judging","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Aleksandar Cvejic, Peter Wonka, Zehui Chen, Zhenyu Li","submitted_at":"2026-05-19T10:59:19Z","abstract_excerpt":"Pairwise human preference prediction is central to evaluating code-generation systems, where quality often depends on task-specific trade-offs beyond functional correctness. While rubric-based LLM judges improve interpretability by decomposing evaluation into explicit criteria, most existing pipelines remain pointwise: they score each response independently and derive preferences by comparing aggregated scores. We show that this design is poorly matched to pairwise code preference prediction and can underperform a strong monolithic judge. We propose CriterAlign, a criterion-centric framework t"},"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":"2605.19665","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-05-19T10:59:19Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"22600d7554e6ead514f4be679e7ac3239c57635fc87ed543f60774bb0ae692c1","abstract_canon_sha256":"40d9ea3bebfbb29b55125c5a70bf2cb05d2c692ab5bcb72bad5cf82d84d4aa41"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T01:05:56.856800Z","signature_b64":"lw5Na+1g/4OPvj61RX7Ypoek6zI/eKQbdJkXh+JVxtdT3AwiekhKPKJGSjem5o0ngycjZhjuDYUQIq3aQXffAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e2e7888ba9a0ce3648c795ed347023c88985f8e7ef51fcd84fb7d693101b108f","last_reissued_at":"2026-05-20T01:05:56.856229Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T01:05:56.856229Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CriterAlign: Criterion-Centric Rationale Alignment for Code Preference Judging","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Aleksandar Cvejic, Peter Wonka, Zehui Chen, Zhenyu Li","submitted_at":"2026-05-19T10:59:19Z","abstract_excerpt":"Pairwise human preference prediction is central to evaluating code-generation systems, where quality often depends on task-specific trade-offs beyond functional correctness. While rubric-based LLM judges improve interpretability by decomposing evaluation into explicit criteria, most existing pipelines remain pointwise: they score each response independently and derive preferences by comparing aggregated scores. We show that this design is poorly matched to pairwise code preference prediction and can underperform a strong monolithic judge. We propose CriterAlign, a criterion-centric framework t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.19665","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/2605.19665/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":"2605.19665","created_at":"2026-05-20T01:05:56.856299+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.19665v1","created_at":"2026-05-20T01:05:56.856299+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.19665","created_at":"2026-05-20T01:05:56.856299+00:00"},{"alias_kind":"pith_short_12","alias_value":"4LTYRC5JUDHD","created_at":"2026-05-20T01:05:56.856299+00:00"},{"alias_kind":"pith_short_16","alias_value":"4LTYRC5JUDHDMSGH","created_at":"2026-05-20T01:05:56.856299+00:00"},{"alias_kind":"pith_short_8","alias_value":"4LTYRC5J","created_at":"2026-05-20T01:05:56.856299+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/4LTYRC5JUDHDMSGHSXWTI4BDZC","json":"https://pith.science/pith/4LTYRC5JUDHDMSGHSXWTI4BDZC.json","graph_json":"https://pith.science/api/pith-number/4LTYRC5JUDHDMSGHSXWTI4BDZC/graph.json","events_json":"https://pith.science/api/pith-number/4LTYRC5JUDHDMSGHSXWTI4BDZC/events.json","paper":"https://pith.science/paper/4LTYRC5J"},"agent_actions":{"view_html":"https://pith.science/pith/4LTYRC5JUDHDMSGHSXWTI4BDZC","download_json":"https://pith.science/pith/4LTYRC5JUDHDMSGHSXWTI4BDZC.json","view_paper":"https://pith.science/paper/4LTYRC5J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.19665&json=true","fetch_graph":"https://pith.science/api/pith-number/4LTYRC5JUDHDMSGHSXWTI4BDZC/graph.json","fetch_events":"https://pith.science/api/pith-number/4LTYRC5JUDHDMSGHSXWTI4BDZC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4LTYRC5JUDHDMSGHSXWTI4BDZC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4LTYRC5JUDHDMSGHSXWTI4BDZC/action/storage_attestation","attest_author":"https://pith.science/pith/4LTYRC5JUDHDMSGHSXWTI4BDZC/action/author_attestation","sign_citation":"https://pith.science/pith/4LTYRC5JUDHDMSGHSXWTI4BDZC/action/citation_signature","submit_replication":"https://pith.science/pith/4LTYRC5JUDHDMSGHSXWTI4BDZC/action/replication_record"}},"created_at":"2026-05-20T01:05:56.856299+00:00","updated_at":"2026-05-20T01:05:56.856299+00:00"}