{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KORTA25IQMY64FPLUQSSGRKB3O","short_pith_number":"pith:KORTA25I","schema_version":"1.0","canonical_sha256":"53a3306ba88331ee15eba425234541dba04c881984753f9e45c98730f9b54728","source":{"kind":"arxiv","id":"2606.29296","version":1},"attestation_state":"computed","paper":{"title":"Process Advantage Signal Shaping: A Paradigm-Agnostic Middleware for Process-Supervised RL in LLM Reasoners","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Chao Wang, Hongtao Tian, Tao Yang, Ting Yao, Wenbo Ding, Yunsheng Shi","submitted_at":"2026-06-28T09:36:43Z","abstract_excerpt":"Group Relative Policy Optimization (GRPO) is a default recipe for process-supervised reinforcement learning of LLM reasoners, and dense process supervision -- via learned process reward models (PRMs) or on-policy-distillation KL signals -- is a common way to densify its otherwise weak outcome reward. Layering such a step-level signal on top of GRPO's group-standardized advantage, however, exposes three structural pathologies: \\emph{channel contamination} between the pooled process, outcome, and format streams at group standardization; \\emph{resolution mismatch} between the granularity of the p"},"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":"2606.29296","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-28T09:36:43Z","cross_cats_sorted":[],"title_canon_sha256":"0ae0bda9661b5f8e08a0faa3cc4630bc7e4f808e485ad302094fd6d5d92f2b36","abstract_canon_sha256":"9f1261590861dcaada965ebcc332b59d0f8f6787a99a3fe444ef5742b0602f42"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-30T01:18:00.617358Z","signature_b64":"jXpu9i832zsSTItmqr/SkuIXRoCRkMFvXdVcEtEw/9sqCRiuAvGCNd0133j87AZDqWbAtbDBkzeGecM5cgLWCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"53a3306ba88331ee15eba425234541dba04c881984753f9e45c98730f9b54728","last_reissued_at":"2026-06-30T01:18:00.616833Z","signature_status":"signed_v1","first_computed_at":"2026-06-30T01:18:00.616833Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Process Advantage Signal Shaping: A Paradigm-Agnostic Middleware for Process-Supervised RL in LLM Reasoners","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Chao Wang, Hongtao Tian, Tao Yang, Ting Yao, Wenbo Ding, Yunsheng Shi","submitted_at":"2026-06-28T09:36:43Z","abstract_excerpt":"Group Relative Policy Optimization (GRPO) is a default recipe for process-supervised reinforcement learning of LLM reasoners, and dense process supervision -- via learned process reward models (PRMs) or on-policy-distillation KL signals -- is a common way to densify its otherwise weak outcome reward. Layering such a step-level signal on top of GRPO's group-standardized advantage, however, exposes three structural pathologies: \\emph{channel contamination} between the pooled process, outcome, and format streams at group standardization; \\emph{resolution mismatch} between the granularity of the p"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.29296","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/2606.29296/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":"2606.29296","created_at":"2026-06-30T01:18:00.616915+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.29296v1","created_at":"2026-06-30T01:18:00.616915+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.29296","created_at":"2026-06-30T01:18:00.616915+00:00"},{"alias_kind":"pith_short_12","alias_value":"KORTA25IQMY6","created_at":"2026-06-30T01:18:00.616915+00:00"},{"alias_kind":"pith_short_16","alias_value":"KORTA25IQMY64FPL","created_at":"2026-06-30T01:18:00.616915+00:00"},{"alias_kind":"pith_short_8","alias_value":"KORTA25I","created_at":"2026-06-30T01:18:00.616915+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/KORTA25IQMY64FPLUQSSGRKB3O","json":"https://pith.science/pith/KORTA25IQMY64FPLUQSSGRKB3O.json","graph_json":"https://pith.science/api/pith-number/KORTA25IQMY64FPLUQSSGRKB3O/graph.json","events_json":"https://pith.science/api/pith-number/KORTA25IQMY64FPLUQSSGRKB3O/events.json","paper":"https://pith.science/paper/KORTA25I"},"agent_actions":{"view_html":"https://pith.science/pith/KORTA25IQMY64FPLUQSSGRKB3O","download_json":"https://pith.science/pith/KORTA25IQMY64FPLUQSSGRKB3O.json","view_paper":"https://pith.science/paper/KORTA25I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.29296&json=true","fetch_graph":"https://pith.science/api/pith-number/KORTA25IQMY64FPLUQSSGRKB3O/graph.json","fetch_events":"https://pith.science/api/pith-number/KORTA25IQMY64FPLUQSSGRKB3O/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KORTA25IQMY64FPLUQSSGRKB3O/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KORTA25IQMY64FPLUQSSGRKB3O/action/storage_attestation","attest_author":"https://pith.science/pith/KORTA25IQMY64FPLUQSSGRKB3O/action/author_attestation","sign_citation":"https://pith.science/pith/KORTA25IQMY64FPLUQSSGRKB3O/action/citation_signature","submit_replication":"https://pith.science/pith/KORTA25IQMY64FPLUQSSGRKB3O/action/replication_record"}},"created_at":"2026-06-30T01:18:00.616915+00:00","updated_at":"2026-06-30T01:18:00.616915+00:00"}