{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4GGJLRX2LNUQT7S6MXO656EMAG","short_pith_number":"pith:4GGJLRX2","schema_version":"1.0","canonical_sha256":"e18c95c6fa5b6909fe5e65ddeef88c0186fc770630dc8d1654e994cf4a4cec35","source":{"kind":"arxiv","id":"2606.07040","version":1},"attestation_state":"computed","paper":{"title":"Beyond Rubrics: Exploration-Guided Evaluation Skills for Reward Modeling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Daoxin Zhang, Linjuan Wu, Weiming Lu, Xing Yue, Yongliang Shen","submitted_at":"2026-06-05T08:34:06Z","abstract_excerpt":"Open-ended reward modeling requires judges that can follow subtle, domain-specific preferences when verifiable answers are unavailable. Existing rubric-based methods often address this by generating criteria online for each query, but the extra generation step can add inference overhead and produce rigid or misaligned guidance. We introduce Eval-Skill, an exploration-guided method that synthesizes reusable evaluation skills for reward modeling and reframes reward guidance as context evolution rather than parameter training or per-query rubric generation. Using only 100 cases per domain for ski"},"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":"2606.07040","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-06-05T08:34:06Z","cross_cats_sorted":[],"title_canon_sha256":"1da676430a654eb0bff2255673c4253f2b91d24675b19390a19e5b7d40fb2077","abstract_canon_sha256":"f3c6d12122953512d9882ff554ba5c54b502e3d7241826dffbda946a89bc03a6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-08T01:04:42.861023Z","signature_b64":"GuPVx06EMOh8lyKvM/b90U1kSF7s79b1B4HTzsOvofx/pEE/KOX365Ns46mebK//TfzUESj5fkVz+s0ZLr+fAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e18c95c6fa5b6909fe5e65ddeef88c0186fc770630dc8d1654e994cf4a4cec35","last_reissued_at":"2026-06-08T01:04:42.860273Z","signature_status":"signed_v1","first_computed_at":"2026-06-08T01:04:42.860273Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond Rubrics: Exploration-Guided Evaluation Skills for Reward Modeling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Daoxin Zhang, Linjuan Wu, Weiming Lu, Xing Yue, Yongliang Shen","submitted_at":"2026-06-05T08:34:06Z","abstract_excerpt":"Open-ended reward modeling requires judges that can follow subtle, domain-specific preferences when verifiable answers are unavailable. Existing rubric-based methods often address this by generating criteria online for each query, but the extra generation step can add inference overhead and produce rigid or misaligned guidance. We introduce Eval-Skill, an exploration-guided method that synthesizes reusable evaluation skills for reward modeling and reframes reward guidance as context evolution rather than parameter training or per-query rubric generation. Using only 100 cases per domain for ski"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07040","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/2606.07040/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":"2606.07040","created_at":"2026-06-08T01:04:42.860409+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.07040v1","created_at":"2026-06-08T01:04:42.860409+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07040","created_at":"2026-06-08T01:04:42.860409+00:00"},{"alias_kind":"pith_short_12","alias_value":"4GGJLRX2LNUQ","created_at":"2026-06-08T01:04:42.860409+00:00"},{"alias_kind":"pith_short_16","alias_value":"4GGJLRX2LNUQT7S6","created_at":"2026-06-08T01:04:42.860409+00:00"},{"alias_kind":"pith_short_8","alias_value":"4GGJLRX2","created_at":"2026-06-08T01:04:42.860409+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/4GGJLRX2LNUQT7S6MXO656EMAG","json":"https://pith.science/pith/4GGJLRX2LNUQT7S6MXO656EMAG.json","graph_json":"https://pith.science/api/pith-number/4GGJLRX2LNUQT7S6MXO656EMAG/graph.json","events_json":"https://pith.science/api/pith-number/4GGJLRX2LNUQT7S6MXO656EMAG/events.json","paper":"https://pith.science/paper/4GGJLRX2"},"agent_actions":{"view_html":"https://pith.science/pith/4GGJLRX2LNUQT7S6MXO656EMAG","download_json":"https://pith.science/pith/4GGJLRX2LNUQT7S6MXO656EMAG.json","view_paper":"https://pith.science/paper/4GGJLRX2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.07040&json=true","fetch_graph":"https://pith.science/api/pith-number/4GGJLRX2LNUQT7S6MXO656EMAG/graph.json","fetch_events":"https://pith.science/api/pith-number/4GGJLRX2LNUQT7S6MXO656EMAG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4GGJLRX2LNUQT7S6MXO656EMAG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4GGJLRX2LNUQT7S6MXO656EMAG/action/storage_attestation","attest_author":"https://pith.science/pith/4GGJLRX2LNUQT7S6MXO656EMAG/action/author_attestation","sign_citation":"https://pith.science/pith/4GGJLRX2LNUQT7S6MXO656EMAG/action/citation_signature","submit_replication":"https://pith.science/pith/4GGJLRX2LNUQT7S6MXO656EMAG/action/replication_record"}},"created_at":"2026-06-08T01:04:42.860409+00:00","updated_at":"2026-06-08T01:04:42.860409+00:00"}