{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:LQJIKFIBAABI5MW4EPXG35EBE6","short_pith_number":"pith:LQJIKFIB","schema_version":"1.0","canonical_sha256":"5c1285150100028eb2dc23ee6df48127b4e4088ba1d03a4809913fa9d0c1c514","source":{"kind":"arxiv","id":"2601.08654","version":2},"attestation_state":"computed","paper":{"title":"From Rubrics to Reliable Scores: Evidence-Grounded Text Evaluation with LLM Judges","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bolin Shen, Huaiyuan Yao, Hua Wei, Wanpeng Xu, Yihan Hong, Yushun Dong","submitted_at":"2026-01-13T15:31:42Z","abstract_excerpt":"Rubric-based text evaluation increasingly uses large language models (LLMs) as scalable judges, but aligning frozen black-box models with human scoring standards remains challenging. We formulate this challenge as a criteria-transfer problem: the goal is not merely to prompt an LLM to assign a score, but to transfer human rubric intent into a stable, auditable, and human-aligned scoring protocol. We identify three recurring failure modes in LLM-based rubric scoring: rubric execution drift, unverifiable score attribution, and human-scale misalignment. To address these failure modes, we introduc"},"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":"2601.08654","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2026-01-13T15:31:42Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"da23c475d7d3238b95daef93fc96d3b004dbb18a01703d86fa44d07fe32f3719","abstract_canon_sha256":"b973b0370e81eaba5945dba11167593162a51412868e0f17340b5e8f78697813"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T00:04:12.738726Z","signature_b64":"HmX522MdV5lW54ByUepLDWxImXmQSDTKVBJ1rtJBRCcH9kbk2EmzbEkcRTpntaR4PstFkX3wUcIl5qBXE7T1Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5c1285150100028eb2dc23ee6df48127b4e4088ba1d03a4809913fa9d0c1c514","last_reissued_at":"2026-05-29T00:04:12.738007Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T00:04:12.738007Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Rubrics to Reliable Scores: Evidence-Grounded Text Evaluation with LLM Judges","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bolin Shen, Huaiyuan Yao, Hua Wei, Wanpeng Xu, Yihan Hong, Yushun Dong","submitted_at":"2026-01-13T15:31:42Z","abstract_excerpt":"Rubric-based text evaluation increasingly uses large language models (LLMs) as scalable judges, but aligning frozen black-box models with human scoring standards remains challenging. We formulate this challenge as a criteria-transfer problem: the goal is not merely to prompt an LLM to assign a score, but to transfer human rubric intent into a stable, auditable, and human-aligned scoring protocol. We identify three recurring failure modes in LLM-based rubric scoring: rubric execution drift, unverifiable score attribution, and human-scale misalignment. To address these failure modes, we introduc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.08654","kind":"arxiv","version":2},"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/2601.08654/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":"2601.08654","created_at":"2026-05-29T00:04:12.738082+00:00"},{"alias_kind":"arxiv_version","alias_value":"2601.08654v2","created_at":"2026-05-29T00:04:12.738082+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.08654","created_at":"2026-05-29T00:04:12.738082+00:00"},{"alias_kind":"pith_short_12","alias_value":"LQJIKFIBAABI","created_at":"2026-05-29T00:04:12.738082+00:00"},{"alias_kind":"pith_short_16","alias_value":"LQJIKFIBAABI5MW4","created_at":"2026-05-29T00:04:12.738082+00:00"},{"alias_kind":"pith_short_8","alias_value":"LQJIKFIB","created_at":"2026-05-29T00:04:12.738082+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2604.24074","citing_title":"How Sensitive Are Safety Benchmarks to Judge Configuration Choices?","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06161","citing_title":"Beyond Accuracy: Policy Invariance as a Reliability Test for LLM Safety Judges","ref_index":18,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LQJIKFIBAABI5MW4EPXG35EBE6","json":"https://pith.science/pith/LQJIKFIBAABI5MW4EPXG35EBE6.json","graph_json":"https://pith.science/api/pith-number/LQJIKFIBAABI5MW4EPXG35EBE6/graph.json","events_json":"https://pith.science/api/pith-number/LQJIKFIBAABI5MW4EPXG35EBE6/events.json","paper":"https://pith.science/paper/LQJIKFIB"},"agent_actions":{"view_html":"https://pith.science/pith/LQJIKFIBAABI5MW4EPXG35EBE6","download_json":"https://pith.science/pith/LQJIKFIBAABI5MW4EPXG35EBE6.json","view_paper":"https://pith.science/paper/LQJIKFIB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2601.08654&json=true","fetch_graph":"https://pith.science/api/pith-number/LQJIKFIBAABI5MW4EPXG35EBE6/graph.json","fetch_events":"https://pith.science/api/pith-number/LQJIKFIBAABI5MW4EPXG35EBE6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LQJIKFIBAABI5MW4EPXG35EBE6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LQJIKFIBAABI5MW4EPXG35EBE6/action/storage_attestation","attest_author":"https://pith.science/pith/LQJIKFIBAABI5MW4EPXG35EBE6/action/author_attestation","sign_citation":"https://pith.science/pith/LQJIKFIBAABI5MW4EPXG35EBE6/action/citation_signature","submit_replication":"https://pith.science/pith/LQJIKFIBAABI5MW4EPXG35EBE6/action/replication_record"}},"created_at":"2026-05-29T00:04:12.738082+00:00","updated_at":"2026-05-29T00:04:12.738082+00:00"}