{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:G7RBLVKPEHLWITQNDHOVOSJTGW","short_pith_number":"pith:G7RBLVKP","schema_version":"1.0","canonical_sha256":"37e215d54f21d7644e0d19dd57493335a79935066800659f8fddadfcad82c5ad","source":{"kind":"arxiv","id":"2606.03980","version":1},"attestation_state":"computed","paper":{"title":"Skill-RM: Unifying Heterogeneous Evaluation Criteria via Agent Skill","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Gangwei Jiang, Guanjun Jiang, Jiaqi Guo, Jingwei Ni, Junling Liu, Kai Tang, Mengyu Zhou, Pengyu Cheng, Qinliang su, Siyuan Huang, Tao Chen, Xiaoxi Jiang, Yihao Liu","submitted_at":"2026-06-02T17:56:57Z","abstract_excerpt":"Reward models (RMs) provide critical feedback signals for LLM post-training, notably in reinforced fine-tuning (RFT) and reinforcement learning (RL) pipelines. However, current reward evaluation relies on heterogeneous criteria such as rule-based verifiers, ground-truth references, procedural checklists, and complex rubrics, where a unified mechanism to integrate all types of evidence remains unexplored. To this end, we propose Skill Reward Model (Skill-RM), a unified framework that reformulates reward modeling as the execution of a reusable Reward-Evaluation Skill. By treating reward computat"},"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":"2606.03980","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-02T17:56:57Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"ec8bce7ee9ab4fad9b44b0cbe6afa39ce78feb4e2eb2e199f3ecf8652b237dba","abstract_canon_sha256":"54639d5e1d5b97e652512d761f4cfbe8a9a6b7466e660b324a2820c9ef7eb221"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T02:06:08.571189Z","signature_b64":"KUSwIpZiQxIOCk7ETrfSo+uCUPHj6nZPF9aAnO1aU7g46rjxX2+lnqo9FC338JNLxYl6OqdUTKMGrgAnoKhHAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"37e215d54f21d7644e0d19dd57493335a79935066800659f8fddadfcad82c5ad","last_reissued_at":"2026-06-03T02:06:08.570786Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T02:06:08.570786Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Skill-RM: Unifying Heterogeneous Evaluation Criteria via Agent Skill","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Gangwei Jiang, Guanjun Jiang, Jiaqi Guo, Jingwei Ni, Junling Liu, Kai Tang, Mengyu Zhou, Pengyu Cheng, Qinliang su, Siyuan Huang, Tao Chen, Xiaoxi Jiang, Yihao Liu","submitted_at":"2026-06-02T17:56:57Z","abstract_excerpt":"Reward models (RMs) provide critical feedback signals for LLM post-training, notably in reinforced fine-tuning (RFT) and reinforcement learning (RL) pipelines. However, current reward evaluation relies on heterogeneous criteria such as rule-based verifiers, ground-truth references, procedural checklists, and complex rubrics, where a unified mechanism to integrate all types of evidence remains unexplored. To this end, we propose Skill Reward Model (Skill-RM), a unified framework that reformulates reward modeling as the execution of a reusable Reward-Evaluation Skill. By treating reward computat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.03980","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/2606.03980/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":"2606.03980","created_at":"2026-06-03T02:06:08.570848+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.03980v1","created_at":"2026-06-03T02:06:08.570848+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.03980","created_at":"2026-06-03T02:06:08.570848+00:00"},{"alias_kind":"pith_short_12","alias_value":"G7RBLVKPEHLW","created_at":"2026-06-03T02:06:08.570848+00:00"},{"alias_kind":"pith_short_16","alias_value":"G7RBLVKPEHLWITQN","created_at":"2026-06-03T02:06:08.570848+00:00"},{"alias_kind":"pith_short_8","alias_value":"G7RBLVKP","created_at":"2026-06-03T02:06:08.570848+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/G7RBLVKPEHLWITQNDHOVOSJTGW","json":"https://pith.science/pith/G7RBLVKPEHLWITQNDHOVOSJTGW.json","graph_json":"https://pith.science/api/pith-number/G7RBLVKPEHLWITQNDHOVOSJTGW/graph.json","events_json":"https://pith.science/api/pith-number/G7RBLVKPEHLWITQNDHOVOSJTGW/events.json","paper":"https://pith.science/paper/G7RBLVKP"},"agent_actions":{"view_html":"https://pith.science/pith/G7RBLVKPEHLWITQNDHOVOSJTGW","download_json":"https://pith.science/pith/G7RBLVKPEHLWITQNDHOVOSJTGW.json","view_paper":"https://pith.science/paper/G7RBLVKP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.03980&json=true","fetch_graph":"https://pith.science/api/pith-number/G7RBLVKPEHLWITQNDHOVOSJTGW/graph.json","fetch_events":"https://pith.science/api/pith-number/G7RBLVKPEHLWITQNDHOVOSJTGW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/G7RBLVKPEHLWITQNDHOVOSJTGW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/G7RBLVKPEHLWITQNDHOVOSJTGW/action/storage_attestation","attest_author":"https://pith.science/pith/G7RBLVKPEHLWITQNDHOVOSJTGW/action/author_attestation","sign_citation":"https://pith.science/pith/G7RBLVKPEHLWITQNDHOVOSJTGW/action/citation_signature","submit_replication":"https://pith.science/pith/G7RBLVKPEHLWITQNDHOVOSJTGW/action/replication_record"}},"created_at":"2026-06-03T02:06:08.570848+00:00","updated_at":"2026-06-03T02:06:08.570848+00:00"}