{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:KDEMV4O4TER5VKPFSTFJ4NFSNA","short_pith_number":"pith:KDEMV4O4","canonical_record":{"source":{"id":"2605.20061","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-19T16:19:29Z","cross_cats_sorted":[],"title_canon_sha256":"aaaef1e6a57bf70f83d3f93a4a7aada4e52c5c70bfeeae6bbf1645fb77710377","abstract_canon_sha256":"9db88ca57ced0bd531d3b115f1b53fec14198bfe6a2799cc5c1c9bae817b0e2c"},"schema_version":"1.0"},"canonical_sha256":"50c8caf1dc9923daa9e594ca9e34b26804eb1680c5defded30e91f7cb34d3414","source":{"kind":"arxiv","id":"2605.20061","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.20061","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"arxiv_version","alias_value":"2605.20061v1","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20061","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"pith_short_12","alias_value":"KDEMV4O4TER5","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"pith_short_16","alias_value":"KDEMV4O4TER5VKPF","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"pith_short_8","alias_value":"KDEMV4O4","created_at":"2026-05-20T02:06:00Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:KDEMV4O4TER5VKPFSTFJ4NFSNA","target":"record","payload":{"canonical_record":{"source":{"id":"2605.20061","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-19T16:19:29Z","cross_cats_sorted":[],"title_canon_sha256":"aaaef1e6a57bf70f83d3f93a4a7aada4e52c5c70bfeeae6bbf1645fb77710377","abstract_canon_sha256":"9db88ca57ced0bd531d3b115f1b53fec14198bfe6a2799cc5c1c9bae817b0e2c"},"schema_version":"1.0"},"canonical_sha256":"50c8caf1dc9923daa9e594ca9e34b26804eb1680c5defded30e91f7cb34d3414","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T02:06:00.650422Z","signature_b64":"IVQNYITs4bfZpB5OjAgDD6XaCzVy8oKdDGepjOhTZxSGPbPqvVg+x7vDuZgrPRmuu70dc0L7TdAF0EjZhQEvCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"50c8caf1dc9923daa9e594ca9e34b26804eb1680c5defded30e91f7cb34d3414","last_reissued_at":"2026-05-20T02:06:00.649842Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T02:06:00.649842Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.20061","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":"2026-05-20T02:06:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8v7+ncxUZmPBWoPnRZWgqKlYQcry74iI6XXccAMduajEgR4IwoGLulUGEb0rcY8jyOXTHUiBuh33UU0j6FhmAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-20T16:21:58.441617Z"},"content_sha256":"40a0904744a139cb9636010c74e8a6379f44c03c07ae732a1c5f1103e3817613","schema_version":"1.0","event_id":"sha256:40a0904744a139cb9636010c74e8a6379f44c03c07ae732a1c5f1103e3817613"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:KDEMV4O4TER5VKPFSTFJ4NFSNA","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Rewarding Beliefs, Not Actions: Consistency-Guided Credit Assignment for Long-Horizon Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Liquan Xiao, Minne Li, Sijie Huang, Wenjie Tang, Yuan Zhou","submitted_at":"2026-05-19T16:19:29Z","abstract_excerpt":"Reinforcement learning from verifiable rewards (RLVR) is a promising paradigm for improving large language model (LLM) agents on long-horizon interactive tasks. However, in partially observable environments, incomplete observations cause agent beliefs to drift over time, while delayed rewards obscure the causal impact of intermediate decisions, exacerbating temporal credit assignment challenges. To address this, we propose ReBel (Reward Belief), a process-level reinforcement learning algorithm that explicitly models structured belief states to summarize interaction history and guide subsequent"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20061","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/2605.20061/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":"2026-05-20T02:06:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KHArnQGnfhRh+WLwB5WWgBsxR3nz7RlTlatNMbzmhD30c3RATaFMw4BXu0CaXMZdG1a/cXXt1qxkYVIJWgr2Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-20T16:21:58.441990Z"},"content_sha256":"ead51c55a92d8724aa443be87f2d01bc3682bae38371a91d8c823ae8ca5dc225","schema_version":"1.0","event_id":"sha256:ead51c55a92d8724aa443be87f2d01bc3682bae38371a91d8c823ae8ca5dc225"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:KDEMV4O4TER5VKPFSTFJ4NFSNA","target":"integrity","payload":{"note":"URL 'https://arxiv.org/abs/2502' returned status 404 (Not Found) at last check.","snippet":null,"arxiv_id":"2605.20061","detector":"external_links","evidence":{"url":"https://arxiv.org/abs/2502","final_url":"https://arxiv.org/abs/2502","host_kind":"arxiv","status_code":404,"status_text":"Not Found","verdict_class":"incontrovertible","checked_at_unix":1779276755.3770869},"severity":"advisory","ref_index":null,"audited_at":"2026-05-20T11:32:35.926826Z","event_type":"pith.integrity.v1","detected_doi":null,"detector_url":"https://pith.science/pith-integrity-protocol#external_links","external_url":"https://arxiv.org/abs/2502","finding_type":"dead_url","evidence_hash":"1a6fa2efc33066cba8d6595db6099dee19351c03a94a1bf3f4df28dd6bb86329","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.0.0","detected_arxiv_id":null,"integrity_event_id":5057,"payload_sha256":"5051176ff914f06c1cf290f37990809f46910a21885a8450aeb2aa2b9c696622","signature_b64":"qFTtOYbuiSVW1t6l/i3vFlQVcTOf3VcuM2Hju1v0+hPMrYUDJbzebqzWnoYmjfbSiPYEfN0ZNRoo6yd2RprZDg==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T11:32:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eq5kxEzxjW/3RbtJeZLK1Q8ywFf87mN/jdyAF33O0aD/TJSChyvS1aJt9OlSTVQiDb043qYQUlFEDp8ONDs4BA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-20T16:21:58.442754Z"},"content_sha256":"f3561fad9925c244a9ed43497e941ebc694aec1a1c15a73d0be1f1f86b4ebb98","schema_version":"1.0","event_id":"sha256:f3561fad9925c244a9ed43497e941ebc694aec1a1c15a73d0be1f1f86b4ebb98"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KDEMV4O4TER5VKPFSTFJ4NFSNA/bundle.json","state_url":"https://pith.science/pith/KDEMV4O4TER5VKPFSTFJ4NFSNA/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KDEMV4O4TER5VKPFSTFJ4NFSNA/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":"2026-05-20T16:21:58Z","links":{"resolver":"https://pith.science/pith/KDEMV4O4TER5VKPFSTFJ4NFSNA","bundle":"https://pith.science/pith/KDEMV4O4TER5VKPFSTFJ4NFSNA/bundle.json","state":"https://pith.science/pith/KDEMV4O4TER5VKPFSTFJ4NFSNA/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KDEMV4O4TER5VKPFSTFJ4NFSNA/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:KDEMV4O4TER5VKPFSTFJ4NFSNA","merge_version":"pith-open-graph-merge-v1","event_count":3,"valid_event_count":3,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"9db88ca57ced0bd531d3b115f1b53fec14198bfe6a2799cc5c1c9bae817b0e2c","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-19T16:19:29Z","title_canon_sha256":"aaaef1e6a57bf70f83d3f93a4a7aada4e52c5c70bfeeae6bbf1645fb77710377"},"schema_version":"1.0","source":{"id":"2605.20061","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.20061","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"arxiv_version","alias_value":"2605.20061v1","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20061","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"pith_short_12","alias_value":"KDEMV4O4TER5","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"pith_short_16","alias_value":"KDEMV4O4TER5VKPF","created_at":"2026-05-20T02:06:00Z"},{"alias_kind":"pith_short_8","alias_value":"KDEMV4O4","created_at":"2026-05-20T02:06:00Z"}],"graph_snapshots":[{"event_id":"sha256:ead51c55a92d8724aa443be87f2d01bc3682bae38371a91d8c823ae8ca5dc225","target":"graph","created_at":"2026-05-20T02:06:00Z","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/2605.20061/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement learning from verifiable rewards (RLVR) is a promising paradigm for improving large language model (LLM) agents on long-horizon interactive tasks. However, in partially observable environments, incomplete observations cause agent beliefs to drift over time, while delayed rewards obscure the causal impact of intermediate decisions, exacerbating temporal credit assignment challenges. To address this, we propose ReBel (Reward Belief), a process-level reinforcement learning algorithm that explicitly models structured belief states to summarize interaction history and guide subsequent","authors_text":"Liquan Xiao, Minne Li, Sijie Huang, Wenjie Tang, Yuan Zhou","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-19T16:19:29Z","title":"Rewarding Beliefs, Not Actions: Consistency-Guided Credit Assignment for Long-Horizon Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20061","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":"sha256:40a0904744a139cb9636010c74e8a6379f44c03c07ae732a1c5f1103e3817613","target":"record","created_at":"2026-05-20T02:06:00Z","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":"9db88ca57ced0bd531d3b115f1b53fec14198bfe6a2799cc5c1c9bae817b0e2c","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-19T16:19:29Z","title_canon_sha256":"aaaef1e6a57bf70f83d3f93a4a7aada4e52c5c70bfeeae6bbf1645fb77710377"},"schema_version":"1.0","source":{"id":"2605.20061","kind":"arxiv","version":1}},"canonical_sha256":"50c8caf1dc9923daa9e594ca9e34b26804eb1680c5defded30e91f7cb34d3414","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"50c8caf1dc9923daa9e594ca9e34b26804eb1680c5defded30e91f7cb34d3414","first_computed_at":"2026-05-20T02:06:00.649842Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T02:06:00.649842Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"IVQNYITs4bfZpB5OjAgDD6XaCzVy8oKdDGepjOhTZxSGPbPqvVg+x7vDuZgrPRmuu70dc0L7TdAF0EjZhQEvCg==","signature_status":"signed_v1","signed_at":"2026-05-20T02:06:00.650422Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.20061","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:40a0904744a139cb9636010c74e8a6379f44c03c07ae732a1c5f1103e3817613","sha256:ead51c55a92d8724aa443be87f2d01bc3682bae38371a91d8c823ae8ca5dc225","sha256:f3561fad9925c244a9ed43497e941ebc694aec1a1c15a73d0be1f1f86b4ebb98"],"state_sha256":"dab9b25ab087d4798d2b8c336006cd6aa605250e0c1e0f1ec86e7abf95079687"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"C3+HMpZXHqYy7OD2OdUzzQ0Hejxx9dx7FRzomMWmYMCdvZMHpF1sNrmtpZmvrefDUxJUayaMS5G2KFwlm4omBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-20T16:21:58.444663Z","bundle_sha256":"0b6dd5c9cc1f16fa05c354e3adcbe46d03e54f77e1e3f1269757ec633fd08aed"}}