{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:COBT7Y5PW4VMEY7XT5FPMKDQ4I","short_pith_number":"pith:COBT7Y5P","canonical_record":{"source":{"id":"2502.10325","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-14T17:34:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"47ac6aea72fe3a16d136d0af626083ed77472c8070ba4b19d067c98173136357","abstract_canon_sha256":"7ab7ef86fc6303098af974c71d6da94a1b4180a735bd05f1becba03abfe010ee"},"schema_version":"1.0"},"canonical_sha256":"13833fe3afb72ac263f79f4af62870e20d682f796fa87122315ccfcd9f7bca01","source":{"kind":"arxiv","id":"2502.10325","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.10325","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"arxiv_version","alias_value":"2502.10325v1","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.10325","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"pith_short_12","alias_value":"COBT7Y5PW4VM","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"pith_short_16","alias_value":"COBT7Y5PW4VMEY7X","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"pith_short_8","alias_value":"COBT7Y5P","created_at":"2026-07-05T10:14:32Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:COBT7Y5PW4VMEY7XT5FPMKDQ4I","target":"record","payload":{"canonical_record":{"source":{"id":"2502.10325","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-14T17:34:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"47ac6aea72fe3a16d136d0af626083ed77472c8070ba4b19d067c98173136357","abstract_canon_sha256":"7ab7ef86fc6303098af974c71d6da94a1b4180a735bd05f1becba03abfe010ee"},"schema_version":"1.0"},"canonical_sha256":"13833fe3afb72ac263f79f4af62870e20d682f796fa87122315ccfcd9f7bca01","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:14:32.356441Z","signature_b64":"r6BrJxn+ZVU104UZ7ZbTScSHMh93QgM+/fZ2pfJmTL305p+7fFyUkiCwFAcMRK3c52h3MvnJ2hLD5HqZB3CfAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"13833fe3afb72ac263f79f4af62870e20d682f796fa87122315ccfcd9f7bca01","last_reissued_at":"2026-07-05T10:14:32.355934Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:14:32.355934Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.10325","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:14:32Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xEXrndWdgI448C4HulyxEHcSUMh8JEPD2hc0E+k8TqPQsr+e1VOm6ANoUMlAcom6h5T8muTJJKjMUaElHPrdAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T18:35:01.576589Z"},"content_sha256":"9924ad81508355690ca419b6d13ed5cad418499f52716b7fff9667bd3989158f","schema_version":"1.0","event_id":"sha256:9924ad81508355690ca419b6d13ed5cad418499f52716b7fff9667bd3989158f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:COBT7Y5PW4VMEY7XT5FPMKDQ4I","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Process Reward Models for LLM Agents: Practical Framework and Directions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Sanjiban Choudhury","submitted_at":"2025-02-14T17:34:28Z","abstract_excerpt":"We introduce Agent Process Reward Models (AgentPRM), a simple and scalable framework for training LLM agents to continually improve through interactions. AgentPRM follows a lightweight actor-critic paradigm, using Monte Carlo rollouts to compute reward targets and optimize policies. It requires minimal modifications to existing RLHF pipelines, making it easy to integrate at scale. Beyond AgentPRM, we propose InversePRM, which learns process rewards directly from demonstrations without explicit outcome supervision. We also explore key challenges and opportunities, including exploration, process"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.10325","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/2502.10325/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:14:32Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pwBV7Qow8Hqt9fgF3dFTcj5XRsUKb469YP5O3roeublAxe8HK9hw5dQIOyrjqf+VHTzUZ4GsXa8k8o3AggpXAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T18:35:01.577683Z"},"content_sha256":"38fc29cc8a96f3069498485c461bdefbe40e578a8584ee65ee0cf2fa7677935f","schema_version":"1.0","event_id":"sha256:38fc29cc8a96f3069498485c461bdefbe40e578a8584ee65ee0cf2fa7677935f"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/COBT7Y5PW4VMEY7XT5FPMKDQ4I/bundle.json","state_url":"https://pith.science/pith/COBT7Y5PW4VMEY7XT5FPMKDQ4I/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/COBT7Y5PW4VMEY7XT5FPMKDQ4I/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-09T18:35:01Z","links":{"resolver":"https://pith.science/pith/COBT7Y5PW4VMEY7XT5FPMKDQ4I","bundle":"https://pith.science/pith/COBT7Y5PW4VMEY7XT5FPMKDQ4I/bundle.json","state":"https://pith.science/pith/COBT7Y5PW4VMEY7XT5FPMKDQ4I/state.json","well_known_bundle":"https://pith.science/.well-known/pith/COBT7Y5PW4VMEY7XT5FPMKDQ4I/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:COBT7Y5PW4VMEY7XT5FPMKDQ4I","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"7ab7ef86fc6303098af974c71d6da94a1b4180a735bd05f1becba03abfe010ee","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-14T17:34:28Z","title_canon_sha256":"47ac6aea72fe3a16d136d0af626083ed77472c8070ba4b19d067c98173136357"},"schema_version":"1.0","source":{"id":"2502.10325","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.10325","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"arxiv_version","alias_value":"2502.10325v1","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.10325","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"pith_short_12","alias_value":"COBT7Y5PW4VM","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"pith_short_16","alias_value":"COBT7Y5PW4VMEY7X","created_at":"2026-07-05T10:14:32Z"},{"alias_kind":"pith_short_8","alias_value":"COBT7Y5P","created_at":"2026-07-05T10:14:32Z"}],"graph_snapshots":[{"event_id":"sha256:38fc29cc8a96f3069498485c461bdefbe40e578a8584ee65ee0cf2fa7677935f","target":"graph","created_at":"2026-07-05T10:14:32Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2502.10325/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We introduce Agent Process Reward Models (AgentPRM), a simple and scalable framework for training LLM agents to continually improve through interactions. AgentPRM follows a lightweight actor-critic paradigm, using Monte Carlo rollouts to compute reward targets and optimize policies. It requires minimal modifications to existing RLHF pipelines, making it easy to integrate at scale. Beyond AgentPRM, we propose InversePRM, which learns process rewards directly from demonstrations without explicit outcome supervision. We also explore key challenges and opportunities, including exploration, process","authors_text":"Sanjiban Choudhury","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-14T17:34:28Z","title":"Process Reward Models for LLM Agents: Practical Framework and Directions"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.10325","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:9924ad81508355690ca419b6d13ed5cad418499f52716b7fff9667bd3989158f","target":"record","created_at":"2026-07-05T10:14:32Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"7ab7ef86fc6303098af974c71d6da94a1b4180a735bd05f1becba03abfe010ee","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-14T17:34:28Z","title_canon_sha256":"47ac6aea72fe3a16d136d0af626083ed77472c8070ba4b19d067c98173136357"},"schema_version":"1.0","source":{"id":"2502.10325","kind":"arxiv","version":1}},"canonical_sha256":"13833fe3afb72ac263f79f4af62870e20d682f796fa87122315ccfcd9f7bca01","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"13833fe3afb72ac263f79f4af62870e20d682f796fa87122315ccfcd9f7bca01","first_computed_at":"2026-07-05T10:14:32.355934Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:14:32.355934Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"r6BrJxn+ZVU104UZ7ZbTScSHMh93QgM+/fZ2pfJmTL305p+7fFyUkiCwFAcMRK3c52h3MvnJ2hLD5HqZB3CfAw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:14:32.356441Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.10325","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9924ad81508355690ca419b6d13ed5cad418499f52716b7fff9667bd3989158f","sha256:38fc29cc8a96f3069498485c461bdefbe40e578a8584ee65ee0cf2fa7677935f"],"state_sha256":"5b09718f3ada0e3c50877f6d0977f9bce4704c35f4703c1dddf6fc317a736259"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6W2RwnNaD9AvE6OsQ9WijH2qFRE+KW21vbD9DI0V/vydqm6Fp6iw1oN7ptCRNSdqTLZHxpvjhI00Q5CECYXhCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T18:35:01.590950Z","bundle_sha256":"df3db4005d02c60eb01afc1f9035db416d5b678ec4eb8a51cf58aefd84bc6028"}}