{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:YMJNFF3GKDF5ITPVHOUQKO6D7J","short_pith_number":"pith:YMJNFF3G","canonical_record":{"source":{"id":"2604.09482","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-10T16:45:44Z","cross_cats_sorted":[],"title_canon_sha256":"92b491970cb747dc463d80ee2cb8307cf3a866ffd7de9570096323fae0af1e83","abstract_canon_sha256":"3d68182b83fcba80a276f656f3b39509f5e035d32b9d9de46c8f8c9f974d5d13"},"schema_version":"1.0"},"canonical_sha256":"c312d2976650cbd44df53ba9053bc3fa6004c1ba2fd631103acb5c432caaf683","source":{"kind":"arxiv","id":"2604.09482","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.09482","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"arxiv_version","alias_value":"2604.09482v2","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.09482","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"pith_short_12","alias_value":"YMJNFF3GKDF5","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"pith_short_16","alias_value":"YMJNFF3GKDF5ITPV","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"pith_short_8","alias_value":"YMJNFF3G","created_at":"2026-06-02T02:04:52Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:YMJNFF3GKDF5ITPVHOUQKO6D7J","target":"record","payload":{"canonical_record":{"source":{"id":"2604.09482","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-10T16:45:44Z","cross_cats_sorted":[],"title_canon_sha256":"92b491970cb747dc463d80ee2cb8307cf3a866ffd7de9570096323fae0af1e83","abstract_canon_sha256":"3d68182b83fcba80a276f656f3b39509f5e035d32b9d9de46c8f8c9f974d5d13"},"schema_version":"1.0"},"canonical_sha256":"c312d2976650cbd44df53ba9053bc3fa6004c1ba2fd631103acb5c432caaf683","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:52.913585Z","signature_b64":"Jdou72kvOTC3SEDtheQhEtmt3rzIXNnMZwuBiv+eVszjhKUxm4C4L66v1wGgBCK5hJ+V7qzMDeu1e+lZ56X+Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c312d2976650cbd44df53ba9053bc3fa6004c1ba2fd631103acb5c432caaf683","last_reissued_at":"2026-06-02T02:04:52.913127Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:52.913127Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.09482","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-02T02:04:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"T/+ZnWBNTsaPWnj/JzaOW0YerBlVOlsIljCOz4nVFpT60hPPHoIc5ifVJRqWdZas/294Z1zH2ITayQrXjm4/Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-25T11:11:17.265631Z"},"content_sha256":"addece82f21d32107e1b0ff68ccf5d84e206087c8397ec3c4109645412143397","schema_version":"1.0","event_id":"sha256:addece82f21d32107e1b0ff68ccf5d84e206087c8397ec3c4109645412143397"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:YMJNFF3GKDF5ITPVHOUQKO6D7J","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Process Reward Agents for Steering Knowledge-Intensive Reasoning","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"Process Reward Agents supply online step-wise rewards from external knowledge to steer reasoning in frozen language models.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jiwoong Sohn, Kenneth Styppa, Michael Moor, Tomasz Sternal, Torsten Hoefler","submitted_at":"2026-04-10T16:45:44Z","abstract_excerpt":"Reasoning in knowledge-intensive domains remains challenging as intermediate steps are often not locally verifiable: unlike math or code, evaluating step correctness may require synthesizing clues across large external knowledge sources. As a result, subtle errors can propagate through reasoning traces, potentially never to be detected. Prior work has proposed process reward models (PRMs), including retrieval-augmented variants, but these methods operate post hoc, scoring completed trajectories, which prevents their integration into dynamic inference procedures. Here, we introduce Process Rewa"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"PRA consistently outperforms strong baselines, achieving 80.8% accuracy on MedQA with Qwen3-4B, a new state of the art at the 4B scale. Importantly, PRA generalizes to unseen frozen policy models ranging from 0.5B to 8B parameters, improving their accuracy by up to 25.7% without any policy model updates.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That retrieval-augmented process rewards can be computed reliably and cheaply at every generation step from external knowledge sources without introducing undetected errors or prohibitive latency that would negate the search benefits.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Process Reward Agents enable online step-wise guidance for frozen AI models in medical reasoning, raising accuracy to 80.8% on MedQA and up to 25.7% gains across 0.5B-8B models without policy updates.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Process Reward Agents supply online step-wise rewards from external knowledge to steer reasoning in frozen language models.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"bf2bc304333b3d191247118c6c200e1e5447a6369ea4569439463b0c149ac691"},"source":{"id":"2604.09482","kind":"arxiv","version":2},"verdict":{"id":"5aff9e65-e160-4734-9b40-00819bb5c476","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T17:57:44.951995Z","strongest_claim":"PRA consistently outperforms strong baselines, achieving 80.8% accuracy on MedQA with Qwen3-4B, a new state of the art at the 4B scale. Importantly, PRA generalizes to unseen frozen policy models ranging from 0.5B to 8B parameters, improving their accuracy by up to 25.7% without any policy model updates.","one_line_summary":"Process Reward Agents enable online step-wise guidance for frozen AI models in medical reasoning, raising accuracy to 80.8% on MedQA and up to 25.7% gains across 0.5B-8B models without policy updates.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That retrieval-augmented process rewards can be computed reliably and cheaply at every generation step from external knowledge sources without introducing undetected errors or prohibitive latency that would negate the search benefits.","pith_extraction_headline":"Process Reward Agents supply online step-wise rewards from external knowledge to steer reasoning in frozen language models."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.09482/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":"5aff9e65-e160-4734-9b40-00819bb5c476"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-02T02:04:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+5Bq9smd8oElrO1tdFpGJ2VDJ16aQip6GzKFe+JcipeV0U1Dwnxnft7Luh2vADmwcq1eLL8CNENuEQTHIYpIBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-25T11:11:17.266121Z"},"content_sha256":"07b40049b3422333b88eaa3fb138552fd82a17443d59e3048facc1d8852ef533","schema_version":"1.0","event_id":"sha256:07b40049b3422333b88eaa3fb138552fd82a17443d59e3048facc1d8852ef533"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/YMJNFF3GKDF5ITPVHOUQKO6D7J/bundle.json","state_url":"https://pith.science/pith/YMJNFF3GKDF5ITPVHOUQKO6D7J/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/YMJNFF3GKDF5ITPVHOUQKO6D7J/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":"2026-06-25T11:11:17Z","links":{"resolver":"https://pith.science/pith/YMJNFF3GKDF5ITPVHOUQKO6D7J","bundle":"https://pith.science/pith/YMJNFF3GKDF5ITPVHOUQKO6D7J/bundle.json","state":"https://pith.science/pith/YMJNFF3GKDF5ITPVHOUQKO6D7J/state.json","well_known_bundle":"https://pith.science/.well-known/pith/YMJNFF3GKDF5ITPVHOUQKO6D7J/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:YMJNFF3GKDF5ITPVHOUQKO6D7J","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":"3d68182b83fcba80a276f656f3b39509f5e035d32b9d9de46c8f8c9f974d5d13","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-10T16:45:44Z","title_canon_sha256":"92b491970cb747dc463d80ee2cb8307cf3a866ffd7de9570096323fae0af1e83"},"schema_version":"1.0","source":{"id":"2604.09482","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.09482","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"arxiv_version","alias_value":"2604.09482v2","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.09482","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"pith_short_12","alias_value":"YMJNFF3GKDF5","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"pith_short_16","alias_value":"YMJNFF3GKDF5ITPV","created_at":"2026-06-02T02:04:52Z"},{"alias_kind":"pith_short_8","alias_value":"YMJNFF3G","created_at":"2026-06-02T02:04:52Z"}],"graph_snapshots":[{"event_id":"sha256:07b40049b3422333b88eaa3fb138552fd82a17443d59e3048facc1d8852ef533","target":"graph","created_at":"2026-06-02T02:04:52Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"PRA consistently outperforms strong baselines, achieving 80.8% accuracy on MedQA with Qwen3-4B, a new state of the art at the 4B scale. Importantly, PRA generalizes to unseen frozen policy models ranging from 0.5B to 8B parameters, improving their accuracy by up to 25.7% without any policy model updates."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That retrieval-augmented process rewards can be computed reliably and cheaply at every generation step from external knowledge sources without introducing undetected errors or prohibitive latency that would negate the search benefits."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Process Reward Agents enable online step-wise guidance for frozen AI models in medical reasoning, raising accuracy to 80.8% on MedQA and up to 25.7% gains across 0.5B-8B models without policy updates."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Process Reward Agents supply online step-wise rewards from external knowledge to steer reasoning in frozen language models."}],"snapshot_sha256":"bf2bc304333b3d191247118c6c200e1e5447a6369ea4569439463b0c149ac691"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.09482/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reasoning in knowledge-intensive domains remains challenging as intermediate steps are often not locally verifiable: unlike math or code, evaluating step correctness may require synthesizing clues across large external knowledge sources. As a result, subtle errors can propagate through reasoning traces, potentially never to be detected. Prior work has proposed process reward models (PRMs), including retrieval-augmented variants, but these methods operate post hoc, scoring completed trajectories, which prevents their integration into dynamic inference procedures. Here, we introduce Process Rewa","authors_text":"Jiwoong Sohn, Kenneth Styppa, Michael Moor, Tomasz Sternal, Torsten Hoefler","cross_cats":[],"headline":"Process Reward Agents supply online step-wise rewards from external knowledge to steer reasoning in frozen language models.","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-10T16:45:44Z","title":"Process Reward Agents for Steering Knowledge-Intensive Reasoning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.09482","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-10T17:57:44.951995Z","id":"5aff9e65-e160-4734-9b40-00819bb5c476","model_set":{"reader":"grok-4.3"},"one_line_summary":"Process Reward Agents enable online step-wise guidance for frozen AI models in medical reasoning, raising accuracy to 80.8% on MedQA and up to 25.7% gains across 0.5B-8B models without policy updates.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Process Reward Agents supply online step-wise rewards from external knowledge to steer reasoning in frozen language models.","strongest_claim":"PRA consistently outperforms strong baselines, achieving 80.8% accuracy on MedQA with Qwen3-4B, a new state of the art at the 4B scale. Importantly, PRA generalizes to unseen frozen policy models ranging from 0.5B to 8B parameters, improving their accuracy by up to 25.7% without any policy model updates.","weakest_assumption":"That retrieval-augmented process rewards can be computed reliably and cheaply at every generation step from external knowledge sources without introducing undetected errors or prohibitive latency that would negate the search benefits."}},"verdict_id":"5aff9e65-e160-4734-9b40-00819bb5c476"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:addece82f21d32107e1b0ff68ccf5d84e206087c8397ec3c4109645412143397","target":"record","created_at":"2026-06-02T02:04:52Z","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":"3d68182b83fcba80a276f656f3b39509f5e035d32b9d9de46c8f8c9f974d5d13","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-10T16:45:44Z","title_canon_sha256":"92b491970cb747dc463d80ee2cb8307cf3a866ffd7de9570096323fae0af1e83"},"schema_version":"1.0","source":{"id":"2604.09482","kind":"arxiv","version":2}},"canonical_sha256":"c312d2976650cbd44df53ba9053bc3fa6004c1ba2fd631103acb5c432caaf683","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c312d2976650cbd44df53ba9053bc3fa6004c1ba2fd631103acb5c432caaf683","first_computed_at":"2026-06-02T02:04:52.913127Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T02:04:52.913127Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Jdou72kvOTC3SEDtheQhEtmt3rzIXNnMZwuBiv+eVszjhKUxm4C4L66v1wGgBCK5hJ+V7qzMDeu1e+lZ56X+Dg==","signature_status":"signed_v1","signed_at":"2026-06-02T02:04:52.913585Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.09482","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:addece82f21d32107e1b0ff68ccf5d84e206087c8397ec3c4109645412143397","sha256:07b40049b3422333b88eaa3fb138552fd82a17443d59e3048facc1d8852ef533"],"state_sha256":"853abd06e9d8081ce7dc2fb80799e3d62bb46e70f2c75a6f5cd7dc7b780bcfcc"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Fh29GsQkjJTveuuBhiKMRBaHyR2gtBAsPngkr8KfXpSNlgZWt+rDaY20Snt+kABkCGtaWWzUKGmmY3T0rw8eAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-25T11:11:17.268392Z","bundle_sha256":"9a453d5004125171f63a4c0fd0fee6ac2134021f20e1b742e41b801187191814"}}