{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:PGE2VZGHPUN44NAJLGDSDZV7XP","short_pith_number":"pith:PGE2VZGH","canonical_record":{"source":{"id":"2606.04923","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T14:18:23Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"f72845742faa938f63c34813f75bd8006542bb9a7f73436fbe219c19c5b06fee","abstract_canon_sha256":"e66bd9dd42c961e5c4aa19a1599552151069ec823d78618ac1a2fee66e51e705"},"schema_version":"1.0"},"canonical_sha256":"7989aae4c77d1bce3409598721e6bfbbc2d4d2962527ac132d904dfd7bdbe444","source":{"kind":"arxiv","id":"2606.04923","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.04923","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"arxiv_version","alias_value":"2606.04923v1","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.04923","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"pith_short_12","alias_value":"PGE2VZGHPUN4","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"pith_short_16","alias_value":"PGE2VZGHPUN44NAJ","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"pith_short_8","alias_value":"PGE2VZGH","created_at":"2026-06-04T01:09:55Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:PGE2VZGHPUN44NAJLGDSDZV7XP","target":"record","payload":{"canonical_record":{"source":{"id":"2606.04923","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T14:18:23Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"f72845742faa938f63c34813f75bd8006542bb9a7f73436fbe219c19c5b06fee","abstract_canon_sha256":"e66bd9dd42c961e5c4aa19a1599552151069ec823d78618ac1a2fee66e51e705"},"schema_version":"1.0"},"canonical_sha256":"7989aae4c77d1bce3409598721e6bfbbc2d4d2962527ac132d904dfd7bdbe444","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-04T01:09:55.867878Z","signature_b64":"pBXBpQ8jQY1G7AM6eCE9CZ53BLeXK6V875zStLcoEijOOoMDAPuUzaSCDgbNzFivw4+/TYlOPuVAgeVgF52fCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7989aae4c77d1bce3409598721e6bfbbc2d4d2962527ac132d904dfd7bdbe444","last_reissued_at":"2026-06-04T01:09:55.867051Z","signature_status":"signed_v1","first_computed_at":"2026-06-04T01:09:55.867051Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.04923","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":"2026-06-04T01:09:55Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CYMO1SAF/E7KElu4Y53ycv0TlcAfoy0if0jRR5m2908qayKT+3pGWhWCiqw+NpZh9F7Q2S+tZJSWedTxpYINDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T12:45:47.005956Z"},"content_sha256":"cadaca8ae04a6c8c5850551ca267c0a2f93e682fc8afec7de154561d52f03330","schema_version":"1.0","event_id":"sha256:cadaca8ae04a6c8c5850551ca267c0a2f93e682fc8afec7de154561d52f03330"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:PGE2VZGHPUN44NAJLGDSDZV7XP","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Reproducing, Analyzing, and Detecting Reward Hacking in Rubric-Based Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Hao Peng, Juanzi Li, Shuo Hou, Xiaozhi Wang, Xuekang Wang, Zhuoyuan Hao","submitted_at":"2026-06-03T14:18:23Z","abstract_excerpt":"Rubric-based reinforcement learning (RL) uses an LLM-as-a-Judge (LaaJ) to score model outputs according to rubrics as rewards. However, policy models may exploit latent biases in the judge, leading to reward hacking and ineffective or unsafe training outcomes. In real-world rubric-based RL, such hacking behaviors are often subtle and entangled with multiple judge biases, making them difficult to analyze, detect, and mitigate. In this paper, we introduce CHERRL, a controllable hacking environment for rubric-based RL. By injecting known biases into LaaJ, CHERRL enables stable reproduction of rew"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.04923","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/2606.04923/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":"2026-06-04T01:09:55Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xFRimllcu9qjLdk25wTFxA6u9UE95Tdh+s25t8yjdKar3KxcwVhXCdUwFNdjVBq1OCg4CJP1RA3iOGKkeXWmAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T12:45:47.006358Z"},"content_sha256":"5e4812fee4a754421b00d34a9136e25a4605e71fa3afdbafb2ed02e37bc6b86a","schema_version":"1.0","event_id":"sha256:5e4812fee4a754421b00d34a9136e25a4605e71fa3afdbafb2ed02e37bc6b86a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/PGE2VZGHPUN44NAJLGDSDZV7XP/bundle.json","state_url":"https://pith.science/pith/PGE2VZGHPUN44NAJLGDSDZV7XP/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/PGE2VZGHPUN44NAJLGDSDZV7XP/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":"2026-06-04T12:45:47Z","links":{"resolver":"https://pith.science/pith/PGE2VZGHPUN44NAJLGDSDZV7XP","bundle":"https://pith.science/pith/PGE2VZGHPUN44NAJLGDSDZV7XP/bundle.json","state":"https://pith.science/pith/PGE2VZGHPUN44NAJLGDSDZV7XP/state.json","well_known_bundle":"https://pith.science/.well-known/pith/PGE2VZGHPUN44NAJLGDSDZV7XP/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:PGE2VZGHPUN44NAJLGDSDZV7XP","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":"e66bd9dd42c961e5c4aa19a1599552151069ec823d78618ac1a2fee66e51e705","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T14:18:23Z","title_canon_sha256":"f72845742faa938f63c34813f75bd8006542bb9a7f73436fbe219c19c5b06fee"},"schema_version":"1.0","source":{"id":"2606.04923","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.04923","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"arxiv_version","alias_value":"2606.04923v1","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.04923","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"pith_short_12","alias_value":"PGE2VZGHPUN4","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"pith_short_16","alias_value":"PGE2VZGHPUN44NAJ","created_at":"2026-06-04T01:09:55Z"},{"alias_kind":"pith_short_8","alias_value":"PGE2VZGH","created_at":"2026-06-04T01:09:55Z"}],"graph_snapshots":[{"event_id":"sha256:5e4812fee4a754421b00d34a9136e25a4605e71fa3afdbafb2ed02e37bc6b86a","target":"graph","created_at":"2026-06-04T01:09:55Z","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/2606.04923/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Rubric-based reinforcement learning (RL) uses an LLM-as-a-Judge (LaaJ) to score model outputs according to rubrics as rewards. However, policy models may exploit latent biases in the judge, leading to reward hacking and ineffective or unsafe training outcomes. In real-world rubric-based RL, such hacking behaviors are often subtle and entangled with multiple judge biases, making them difficult to analyze, detect, and mitigate. In this paper, we introduce CHERRL, a controllable hacking environment for rubric-based RL. By injecting known biases into LaaJ, CHERRL enables stable reproduction of rew","authors_text":"Hao Peng, Juanzi Li, Shuo Hou, Xiaozhi Wang, Xuekang Wang, Zhuoyuan Hao","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T14:18:23Z","title":"Reproducing, Analyzing, and Detecting Reward Hacking in Rubric-Based Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.04923","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:cadaca8ae04a6c8c5850551ca267c0a2f93e682fc8afec7de154561d52f03330","target":"record","created_at":"2026-06-04T01:09:55Z","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":"e66bd9dd42c961e5c4aa19a1599552151069ec823d78618ac1a2fee66e51e705","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T14:18:23Z","title_canon_sha256":"f72845742faa938f63c34813f75bd8006542bb9a7f73436fbe219c19c5b06fee"},"schema_version":"1.0","source":{"id":"2606.04923","kind":"arxiv","version":1}},"canonical_sha256":"7989aae4c77d1bce3409598721e6bfbbc2d4d2962527ac132d904dfd7bdbe444","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7989aae4c77d1bce3409598721e6bfbbc2d4d2962527ac132d904dfd7bdbe444","first_computed_at":"2026-06-04T01:09:55.867051Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-04T01:09:55.867051Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"pBXBpQ8jQY1G7AM6eCE9CZ53BLeXK6V875zStLcoEijOOoMDAPuUzaSCDgbNzFivw4+/TYlOPuVAgeVgF52fCA==","signature_status":"signed_v1","signed_at":"2026-06-04T01:09:55.867878Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.04923","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:cadaca8ae04a6c8c5850551ca267c0a2f93e682fc8afec7de154561d52f03330","sha256:5e4812fee4a754421b00d34a9136e25a4605e71fa3afdbafb2ed02e37bc6b86a"],"state_sha256":"25040af7491ac470323e1e3c072c672cf7c98a065203d0acafb05b89fab4c407"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/vzEQRcTX8gxByC9hOiYnEsT1uoPiWev/tmpgv94aicOX1Qq1YPw58VyyBe9fhSmB5gBTZX9cfnV2U5F5scyDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-04T12:45:47.008533Z","bundle_sha256":"c9b56939e3147e0b0d15715e051d6bf1ad9e23ae59d249a08159f39e1301e55a"}}