{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:VMR7JXUZWB3NCYKSOJRB6FERAU","short_pith_number":"pith:VMR7JXUZ","canonical_record":{"source":{"id":"2210.10760","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-10-19T17:56:10Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"07e2c9a2767dc179a8f6c9c266122319dfab4f082d67b91af49a5bf04c41ebf6","abstract_canon_sha256":"daf52df9b85c624edd98975eb5cde305082b8b219c6580f280442d4995a9830b"},"schema_version":"1.0"},"canonical_sha256":"ab23f4de99b076d1615272621f14910522749409eb5fdb7af5da3b9e9ab480b4","source":{"kind":"arxiv","id":"2210.10760","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2210.10760","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"arxiv_version","alias_value":"2210.10760v1","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.10760","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"pith_short_12","alias_value":"VMR7JXUZWB3N","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"pith_short_16","alias_value":"VMR7JXUZWB3NCYKS","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"pith_short_8","alias_value":"VMR7JXUZ","created_at":"2026-05-19T08:58:26Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:VMR7JXUZWB3NCYKSOJRB6FERAU","target":"record","payload":{"canonical_record":{"source":{"id":"2210.10760","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-10-19T17:56:10Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"07e2c9a2767dc179a8f6c9c266122319dfab4f082d67b91af49a5bf04c41ebf6","abstract_canon_sha256":"daf52df9b85c624edd98975eb5cde305082b8b219c6580f280442d4995a9830b"},"schema_version":"1.0"},"canonical_sha256":"ab23f4de99b076d1615272621f14910522749409eb5fdb7af5da3b9e9ab480b4","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-19T08:58:26.928533Z","signature_b64":"cU1yngT08AdfoRGF4XhE13fIjOL3MpkdV0jCJ7Dw6fnFXmUAnQy0tXcu8aAj9HxJojU44aRZH1QHyjaVyQ5HAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ab23f4de99b076d1615272621f14910522749409eb5fdb7af5da3b9e9ab480b4","last_reissued_at":"2026-05-19T08:58:26.925773Z","signature_status":"signed_v1","first_computed_at":"2026-05-19T08:58:26.925773Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2210.10760","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":"2026-05-19T08:58:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/hpCNVeh3Z1jrN8SOk7oO4OToeF/YPBXD17WmFudefZs5tWcTbmu5/bNcYUxZi+V7+HrFFZ2q4gdQ8NH6eD9Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T21:03:55.522464Z"},"content_sha256":"c15f5acd930a385132030ca8d8f92faa3ac559627de327a1f12c3dc788d76e5f","schema_version":"1.0","event_id":"sha256:c15f5acd930a385132030ca8d8f92faa3ac559627de327a1f12c3dc788d76e5f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:VMR7JXUZWB3NCYKSOJRB6FERAU","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Scaling Laws for Reward Model Overoptimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Jacob Hilton, John Schulman, Leo Gao","submitted_at":"2022-10-19T17:56:10Z","abstract_excerpt":"In reinforcement learning from human feedback, it is common to optimize against a reward model trained to predict human preferences. Because the reward model is an imperfect proxy, optimizing its value too much can hinder ground truth performance, in accordance with Goodhart's law. This effect has been frequently observed, but not carefully measured due to the expense of collecting human preference data. In this work, we use a synthetic setup in which a fixed \"gold-standard\" reward model plays the role of humans, providing labels used to train a proxy reward model. We study how the gold reward"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.10760","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/2210.10760/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":"2026-05-19T08:58:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7F3qd1BB50eSTNcKp1VDLP2L6wZC4ned0Qs9UFxWbR2XPycZ26lIe9rQ4uzbOOIvLcwAmJcbefg8ronG4UbXBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T21:03:55.522844Z"},"content_sha256":"56af6c3052faea908d8d89306d418625fb8877188fb9d9187f927755c36fcf81","schema_version":"1.0","event_id":"sha256:56af6c3052faea908d8d89306d418625fb8877188fb9d9187f927755c36fcf81"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/VMR7JXUZWB3NCYKSOJRB6FERAU/bundle.json","state_url":"https://pith.science/pith/VMR7JXUZWB3NCYKSOJRB6FERAU/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/VMR7JXUZWB3NCYKSOJRB6FERAU/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":"2026-06-03T21:03:55Z","links":{"resolver":"https://pith.science/pith/VMR7JXUZWB3NCYKSOJRB6FERAU","bundle":"https://pith.science/pith/VMR7JXUZWB3NCYKSOJRB6FERAU/bundle.json","state":"https://pith.science/pith/VMR7JXUZWB3NCYKSOJRB6FERAU/state.json","well_known_bundle":"https://pith.science/.well-known/pith/VMR7JXUZWB3NCYKSOJRB6FERAU/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:VMR7JXUZWB3NCYKSOJRB6FERAU","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":"daf52df9b85c624edd98975eb5cde305082b8b219c6580f280442d4995a9830b","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-10-19T17:56:10Z","title_canon_sha256":"07e2c9a2767dc179a8f6c9c266122319dfab4f082d67b91af49a5bf04c41ebf6"},"schema_version":"1.0","source":{"id":"2210.10760","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2210.10760","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"arxiv_version","alias_value":"2210.10760v1","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.10760","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"pith_short_12","alias_value":"VMR7JXUZWB3N","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"pith_short_16","alias_value":"VMR7JXUZWB3NCYKS","created_at":"2026-05-19T08:58:26Z"},{"alias_kind":"pith_short_8","alias_value":"VMR7JXUZ","created_at":"2026-05-19T08:58:26Z"}],"graph_snapshots":[{"event_id":"sha256:56af6c3052faea908d8d89306d418625fb8877188fb9d9187f927755c36fcf81","target":"graph","created_at":"2026-05-19T08:58:26Z","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/2210.10760/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In reinforcement learning from human feedback, it is common to optimize against a reward model trained to predict human preferences. Because the reward model is an imperfect proxy, optimizing its value too much can hinder ground truth performance, in accordance with Goodhart's law. This effect has been frequently observed, but not carefully measured due to the expense of collecting human preference data. In this work, we use a synthetic setup in which a fixed \"gold-standard\" reward model plays the role of humans, providing labels used to train a proxy reward model. We study how the gold reward","authors_text":"Jacob Hilton, John Schulman, Leo Gao","cross_cats":["stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-10-19T17:56:10Z","title":"Scaling Laws for Reward Model Overoptimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.10760","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:c15f5acd930a385132030ca8d8f92faa3ac559627de327a1f12c3dc788d76e5f","target":"record","created_at":"2026-05-19T08:58:26Z","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":"daf52df9b85c624edd98975eb5cde305082b8b219c6580f280442d4995a9830b","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-10-19T17:56:10Z","title_canon_sha256":"07e2c9a2767dc179a8f6c9c266122319dfab4f082d67b91af49a5bf04c41ebf6"},"schema_version":"1.0","source":{"id":"2210.10760","kind":"arxiv","version":1}},"canonical_sha256":"ab23f4de99b076d1615272621f14910522749409eb5fdb7af5da3b9e9ab480b4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ab23f4de99b076d1615272621f14910522749409eb5fdb7af5da3b9e9ab480b4","first_computed_at":"2026-05-19T08:58:26.925773Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-19T08:58:26.925773Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"cU1yngT08AdfoRGF4XhE13fIjOL3MpkdV0jCJ7Dw6fnFXmUAnQy0tXcu8aAj9HxJojU44aRZH1QHyjaVyQ5HAg==","signature_status":"signed_v1","signed_at":"2026-05-19T08:58:26.928533Z","signed_message":"canonical_sha256_bytes"},"source_id":"2210.10760","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c15f5acd930a385132030ca8d8f92faa3ac559627de327a1f12c3dc788d76e5f","sha256:56af6c3052faea908d8d89306d418625fb8877188fb9d9187f927755c36fcf81"],"state_sha256":"2bfb7a13de1e43834217f927ef0398c541d69344c27ba8684d2502738bda1a9e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7WyXVs90+slJbklKjoGJ9t/H1Hrf441udDA3wiLvTeWUcm8LTRiyFL24EG/YZrz5WqOlu5k0pkuUt7uQziUiAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-03T21:03:55.524858Z","bundle_sha256":"7889fc4d92e6d23e5292a01eab90a58a8e32da7c970026e570157010666bd077"}}