{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4BVJROPAXSUCTRS5KZBKUQADBN","short_pith_number":"pith:4BVJROPA","schema_version":"1.0","canonical_sha256":"e06a98b9e0bca829c65d5642aa40030b57466e5bdfc12b0321aadb8be9a211cf","source":{"kind":"arxiv","id":"2605.12667","version":1},"attestation_state":"computed","paper":{"title":"ODRPO: Ordinal Decompositions of Discrete Rewards for Robust Policy Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Decomposing discrete rewards into ordinal binary indicators isolates evaluation noise and stabilizes policy updates in RLAIF without extra compute.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Fei Wang, Inderjit Dhillon, Nirmal Patel","submitted_at":"2026-05-12T19:17:14Z","abstract_excerpt":"The alignment of Large Language Models (LLMs) utilizes Reinforcement Learning from AI Feedback (RLAIF) for non-verifiable domains such as long-form question answering and open-ended instruction following. These domains often rely on LLM based auto-raters to provide granular, multi-tier discrete rewards (e.g., 1-10 rubrics) that are inherently stochastic due to prompt sensitivity and sampling randomness. We empirically verify the stochasticity of auto-raters that can propagate and corrupt standard advantage estimators like GRPO and MaxRL, as a noisy reward samples can skew normalization statist"},"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":true,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.12667","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-12T19:17:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"587504c2283391984dcacc0c61ac7c4a3ba7a095f7037440aca38a8efea03726","abstract_canon_sha256":"20cf49d7c089d5de6656ff33d2420fcd4c87b9714b87d760d8a9a206e4596a4b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:09:50.305122Z","signature_b64":"QgABlwWLNG1ucfjlj4Ux29YmUqTkTyGzBTmyzOnwmQyQZRHgTHfHp5B3ZQ4Ah44xPaWS1HxLl6V4fHOFHBCpBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e06a98b9e0bca829c65d5642aa40030b57466e5bdfc12b0321aadb8be9a211cf","last_reissued_at":"2026-05-18T03:09:50.304220Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:09:50.304220Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ODRPO: Ordinal Decompositions of Discrete Rewards for Robust Policy Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Decomposing discrete rewards into ordinal binary indicators isolates evaluation noise and stabilizes policy updates in RLAIF without extra compute.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Fei Wang, Inderjit Dhillon, Nirmal Patel","submitted_at":"2026-05-12T19:17:14Z","abstract_excerpt":"The alignment of Large Language Models (LLMs) utilizes Reinforcement Learning from AI Feedback (RLAIF) for non-verifiable domains such as long-form question answering and open-ended instruction following. These domains often rely on LLM based auto-raters to provide granular, multi-tier discrete rewards (e.g., 1-10 rubrics) that are inherently stochastic due to prompt sensitivity and sampling randomness. We empirically verify the stochasticity of auto-raters that can propagate and corrupt standard advantage estimators like GRPO and MaxRL, as a noisy reward samples can skew normalization statist"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"ODRPO achieves robust performance on Qwen2.5-7B and Qwen3-4B models, outperforming baselines with relative improvements of upto 14.8% on FACTS-grounding-v2 and 7.5% on Alpaca-Evals. Critically, these gains are achieved with negligible training-time overhead, as ODRPO requires no additional compute per step compared to standard estimators.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That decomposing discrete rewards into ordinal binary indicators structurally isolates evaluation noise and prevents outlier evaluations from corrupting the global update, as stated in the abstract description of the framework.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"ODRPO decomposes discrete rewards into ordinal binary indicators to compute independent advantages and reduce noise corruption in RLAIF policy optimization.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Decomposing discrete rewards into ordinal binary indicators isolates evaluation noise and stabilizes policy updates in RLAIF without extra compute.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"98a7cb3a2c2621f4d7630879b5e9eb097a0ae8b35271b05abecc38f1c6e4bc6c"},"source":{"id":"2605.12667","kind":"arxiv","version":1},"verdict":{"id":"3589802d-eb98-4d60-a062-08a520eeac62","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T21:13:13.575935Z","strongest_claim":"ODRPO achieves robust performance on Qwen2.5-7B and Qwen3-4B models, outperforming baselines with relative improvements of upto 14.8% on FACTS-grounding-v2 and 7.5% on Alpaca-Evals. Critically, these gains are achieved with negligible training-time overhead, as ODRPO requires no additional compute per step compared to standard estimators.","one_line_summary":"ODRPO decomposes discrete rewards into ordinal binary indicators to compute independent advantages and reduce noise corruption in RLAIF policy optimization.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That decomposing discrete rewards into ordinal binary indicators structurally isolates evaluation noise and prevents outlier evaluations from corrupting the global update, as stated in the abstract description of the framework.","pith_extraction_headline":"Decomposing discrete rewards into ordinal binary indicators isolates evaluation noise and stabilizes policy updates in RLAIF without extra compute."},"references":{"count":41,"sample":[{"doi":"10.1038/s41586-025-09422-z","year":null,"title":"Nature645(8081), 633–638 (2025) https://doi.org/10.1038/s41586-025-09422-z","work_id":"9835b482-5032-4135-93dd-82a066677569","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Enigmata: Scaling Logical Reasoning in Large Language Models with Synthetic Verifiable Puzzles , author=. 2025 , eprint=","work_id":"7ea2960e-3de4-4048-be17-ddc3bb8327a6","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Hansen and Duo Peng and Yuhui Zhang and Alejandro Lozano and Min Woo Sun and Emma Lundberg and Serena Yeung-Levy , year=","work_id":"24cdde52-11e0-4409-8656-4a46c92a8552","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"ViCrit: A Verifiable Reinforcement Learning Proxy Task for Visual Perception in VLMs , author=. 2025 , eprint=","work_id":"08971129-3986-43de-9574-8e903aff78da","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"Examining Reasoning LLMs-as-Judges in Non-Verifiable LLM Post-Training , author=. 2026 , eprint=","work_id":"01f55215-0262-4791-87f3-0a2c1cd0bf25","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":41,"snapshot_sha256":"4a64a2eec651d75f8ad382c46ecf342602c7241135526e5dbb9583aacf526677","internal_anchors":2},"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":"2605.12667","created_at":"2026-05-18T03:09:50.304386+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.12667v1","created_at":"2026-05-18T03:09:50.304386+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12667","created_at":"2026-05-18T03:09:50.304386+00:00"},{"alias_kind":"pith_short_12","alias_value":"4BVJROPAXSUC","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"4BVJROPAXSUCTRS5","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"4BVJROPA","created_at":"2026-05-18T12:33:37.589309+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/4BVJROPAXSUCTRS5KZBKUQADBN","json":"https://pith.science/pith/4BVJROPAXSUCTRS5KZBKUQADBN.json","graph_json":"https://pith.science/api/pith-number/4BVJROPAXSUCTRS5KZBKUQADBN/graph.json","events_json":"https://pith.science/api/pith-number/4BVJROPAXSUCTRS5KZBKUQADBN/events.json","paper":"https://pith.science/paper/4BVJROPA"},"agent_actions":{"view_html":"https://pith.science/pith/4BVJROPAXSUCTRS5KZBKUQADBN","download_json":"https://pith.science/pith/4BVJROPAXSUCTRS5KZBKUQADBN.json","view_paper":"https://pith.science/paper/4BVJROPA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.12667&json=true","fetch_graph":"https://pith.science/api/pith-number/4BVJROPAXSUCTRS5KZBKUQADBN/graph.json","fetch_events":"https://pith.science/api/pith-number/4BVJROPAXSUCTRS5KZBKUQADBN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4BVJROPAXSUCTRS5KZBKUQADBN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4BVJROPAXSUCTRS5KZBKUQADBN/action/storage_attestation","attest_author":"https://pith.science/pith/4BVJROPAXSUCTRS5KZBKUQADBN/action/author_attestation","sign_citation":"https://pith.science/pith/4BVJROPAXSUCTRS5KZBKUQADBN/action/citation_signature","submit_replication":"https://pith.science/pith/4BVJROPAXSUCTRS5KZBKUQADBN/action/replication_record"}},"created_at":"2026-05-18T03:09:50.304386+00:00","updated_at":"2026-05-18T03:09:50.304386+00:00"}