{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:EWU7G3DXDD4NVOGTIEEZJUERTK","short_pith_number":"pith:EWU7G3DX","schema_version":"1.0","canonical_sha256":"25a9f36c7718f8dab8d3410994d0919aacd8c3783656df69625f2b6347379c55","source":{"kind":"arxiv","id":"2410.08146","version":1},"attestation_state":"computed","paper":{"title":"Rewarding Progress: Scaling Automated Process Verifiers for LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Adam Fisch, Alekh Agarwal, Amrith Setlur, Aviral Kumar, Chirag Nagpal, Jacob Eisenstein, Jonathan Berant, Rishabh Agarwal, Xinyang Geng","submitted_at":"2024-10-10T17:31:23Z","abstract_excerpt":"A promising approach for improving reasoning in large language models is to use process reward models (PRMs). PRMs provide feedback at each step of a multi-step reasoning trace, potentially improving credit assignment over outcome reward models (ORMs) that only provide feedback at the final step. However, collecting dense, per-step human labels is not scalable, and training PRMs from automatically-labeled data has thus far led to limited gains. To improve a base policy by running search against a PRM or using it as dense rewards for reinforcement learning (RL), we ask: \"How should we design pr"},"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":"2410.08146","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-10T17:31:23Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"902fdefaf98de11fae6312d25228e4eff2eb2f3e5dcf0c49070aba16f779dc5e","abstract_canon_sha256":"c9e9386f9dec4f30d204c77ef55bd2065b8e6142e5864cd72a70abae377bbdbe"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:36:42.130275Z","signature_b64":"488XLloOuui9IPExjD+RzFg9YoVSsPjBI676G+vfx6Bnw/UMcqHoTNragN/BcSpvvyL0fnR4DBvFcILtaE/dDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"25a9f36c7718f8dab8d3410994d0919aacd8c3783656df69625f2b6347379c55","last_reissued_at":"2026-05-21T01:36:42.127656Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:36:42.127656Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rewarding Progress: Scaling Automated Process Verifiers for LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Adam Fisch, Alekh Agarwal, Amrith Setlur, Aviral Kumar, Chirag Nagpal, Jacob Eisenstein, Jonathan Berant, Rishabh Agarwal, Xinyang Geng","submitted_at":"2024-10-10T17:31:23Z","abstract_excerpt":"A promising approach for improving reasoning in large language models is to use process reward models (PRMs). PRMs provide feedback at each step of a multi-step reasoning trace, potentially improving credit assignment over outcome reward models (ORMs) that only provide feedback at the final step. However, collecting dense, per-step human labels is not scalable, and training PRMs from automatically-labeled data has thus far led to limited gains. To improve a base policy by running search against a PRM or using it as dense rewards for reinforcement learning (RL), we ask: \"How should we design pr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.08146","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/2410.08146/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":"2410.08146","created_at":"2026-05-21T01:36:42.127775+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.08146v1","created_at":"2026-05-21T01:36:42.127775+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.08146","created_at":"2026-05-21T01:36:42.127775+00:00"},{"alias_kind":"pith_short_12","alias_value":"EWU7G3DXDD4N","created_at":"2026-05-21T01:36:42.127775+00:00"},{"alias_kind":"pith_short_16","alias_value":"EWU7G3DXDD4NVOGT","created_at":"2026-05-21T01:36:42.127775+00:00"},{"alias_kind":"pith_short_8","alias_value":"EWU7G3DX","created_at":"2026-05-21T01:36:42.127775+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":20,"internal_anchor_count":20,"sample":[{"citing_arxiv_id":"2605.18851","citing_title":"STRIDE: Learnable Stepwise Language Feedback for LLM Reasoning","ref_index":8,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19436","citing_title":"CEPO: RLVR Self-Distillation using Contrastive Evidence Policy Optimization","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19228","citing_title":"Diagnosing Multi-step Reasoning Failures in Black-box LLMs via Stepwise Confidence Attribution","ref_index":49,"is_internal_anchor":true},{"citing_arxiv_id":"2605.20033","citing_title":"A Nash Equilibrium Framework For Training-Free Multimodal Step Verification","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2508.15202","citing_title":"Fin-PRM: A Domain-Specialized Process Reward Model for Financial Reasoning in Large Language Models","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2505.15134","citing_title":"The Unreasonable Effectiveness of Entropy Minimization in LLM Reasoning","ref_index":68,"is_internal_anchor":true},{"citing_arxiv_id":"2510.14703","citing_title":"ToolPRM: Fine-Grained Inference Scaling of Structured Outputs for Function Calling","ref_index":32,"is_internal_anchor":true},{"citing_arxiv_id":"2601.03555","citing_title":"SCRIBE: Structured Mid-Level Supervision for Tool-Using Language Models","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2510.13786","citing_title":"The Art of Scaling Reinforcement Learning Compute for LLMs","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2501.09686","citing_title":"Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language Models","ref_index":125,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12947","citing_title":"When Should an AI Workflow Release? Always-Valid Inference for Black-Box Generate-Verify Systems","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2502.01456","citing_title":"Process Reinforcement through Implicit Rewards","ref_index":88,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06650","citing_title":"Beyond Negative Rollouts: Positive-Only Policy Optimization with Implicit Negative Gradients","ref_index":37,"is_internal_anchor":true},{"citing_arxiv_id":"2605.02290","citing_title":"Distilling Long-CoT Reasoning through Collaborative Step-wise Multi-Teacher Decoding","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"2604.10660","citing_title":"Efficient Process Reward Modeling via Contrastive Mutual Information","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2604.09107","citing_title":"TensorHub: Scalable and Elastic Weight Transfer for LLM RL Training","ref_index":31,"is_internal_anchor":true},{"citing_arxiv_id":"2604.16453","citing_title":"Sampling for Quality: Training-Free Reward-Guided LLM Decoding via Sequential Monte Carlo","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2604.17282","citing_title":"MedPRMBench: A Fine-grained Benchmark for Process Reward Models in Medical Reasoning","ref_index":26,"is_internal_anchor":true},{"citing_arxiv_id":"2604.20659","citing_title":"GRPO-VPS: Enhancing Group Relative Policy Optimization with Verifiable Process Supervision for Effective Reasoning","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"2605.02395","citing_title":"Controllable and Verifiable Process Data Synthesis for Process Reward Models","ref_index":10,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EWU7G3DXDD4NVOGTIEEZJUERTK","json":"https://pith.science/pith/EWU7G3DXDD4NVOGTIEEZJUERTK.json","graph_json":"https://pith.science/api/pith-number/EWU7G3DXDD4NVOGTIEEZJUERTK/graph.json","events_json":"https://pith.science/api/pith-number/EWU7G3DXDD4NVOGTIEEZJUERTK/events.json","paper":"https://pith.science/paper/EWU7G3DX"},"agent_actions":{"view_html":"https://pith.science/pith/EWU7G3DXDD4NVOGTIEEZJUERTK","download_json":"https://pith.science/pith/EWU7G3DXDD4NVOGTIEEZJUERTK.json","view_paper":"https://pith.science/paper/EWU7G3DX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.08146&json=true","fetch_graph":"https://pith.science/api/pith-number/EWU7G3DXDD4NVOGTIEEZJUERTK/graph.json","fetch_events":"https://pith.science/api/pith-number/EWU7G3DXDD4NVOGTIEEZJUERTK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EWU7G3DXDD4NVOGTIEEZJUERTK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EWU7G3DXDD4NVOGTIEEZJUERTK/action/storage_attestation","attest_author":"https://pith.science/pith/EWU7G3DXDD4NVOGTIEEZJUERTK/action/author_attestation","sign_citation":"https://pith.science/pith/EWU7G3DXDD4NVOGTIEEZJUERTK/action/citation_signature","submit_replication":"https://pith.science/pith/EWU7G3DXDD4NVOGTIEEZJUERTK/action/replication_record"}},"created_at":"2026-05-21T01:36:42.127775+00:00","updated_at":"2026-05-21T01:36:42.127775+00:00"}