{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4P5XPPGT34NNX4JBXIBDBAJTEE","short_pith_number":"pith:4P5XPPGT","schema_version":"1.0","canonical_sha256":"e3fb77bcd3df1adbf121ba023081332135e43686821e54f8e4b9c8adc4485bd2","source":{"kind":"arxiv","id":"2506.01692","version":1},"attestation_state":"computed","paper":{"title":"A Descriptive and Normative Theory of Human Beliefs in RLHF","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Frank Chiu, Scott Niekum, Shripad Deshmukh, Sylee Dandekar, W. Bradley Knox","submitted_at":"2025-06-02T13:52:55Z","abstract_excerpt":"Human preferences in RLHF are typically modeled as a function of the human's reward function or corresponding optimal state-action values. In this work, we propose that human beliefs about the capabilities of the agent being trained also play a key role in preference generation. We examine two questions related to this hypothesis, one descriptive and one normative, respectively: Do human labelers' beliefs about agent capabilities affect the preferences that they provide? And what is the ideal set of beliefs about an agent -- and resulting preferences -- for humans to have? We propose a new pre"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2506.01692","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-06-02T13:52:55Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"6a10ee1a862147779c215bc17d93c904d2a67102d9bedd5baa08742cf9dd9426","abstract_canon_sha256":"addf39ec5673ddb5ee455b4a182d6c94da62873a50ec90ca99c7c1509b12055c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:14:17.971374Z","signature_b64":"hgoPxfNaGcsvHEKURfZc5YeM52j9ZUgDPSfPKMkkEjksmJIDsaMevqSXF9V+Uid1+VaX8xttWTEXlgTnW/NUCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e3fb77bcd3df1adbf121ba023081332135e43686821e54f8e4b9c8adc4485bd2","last_reissued_at":"2026-07-05T11:14:17.970897Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:14:17.970897Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Descriptive and Normative Theory of Human Beliefs in RLHF","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Frank Chiu, Scott Niekum, Shripad Deshmukh, Sylee Dandekar, W. Bradley Knox","submitted_at":"2025-06-02T13:52:55Z","abstract_excerpt":"Human preferences in RLHF are typically modeled as a function of the human's reward function or corresponding optimal state-action values. In this work, we propose that human beliefs about the capabilities of the agent being trained also play a key role in preference generation. We examine two questions related to this hypothesis, one descriptive and one normative, respectively: Do human labelers' beliefs about agent capabilities affect the preferences that they provide? And what is the ideal set of beliefs about an agent -- and resulting preferences -- for humans to have? We propose a new pre"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.01692","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/2506.01692/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2506.01692","created_at":"2026-07-05T11:14:17.970955+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.01692v1","created_at":"2026-07-05T11:14:17.970955+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.01692","created_at":"2026-07-05T11:14:17.970955+00:00"},{"alias_kind":"pith_short_12","alias_value":"4P5XPPGT34NN","created_at":"2026-07-05T11:14:17.970955+00:00"},{"alias_kind":"pith_short_16","alias_value":"4P5XPPGT34NNX4JB","created_at":"2026-07-05T11:14:17.970955+00:00"},{"alias_kind":"pith_short_8","alias_value":"4P5XPPGT","created_at":"2026-07-05T11:14:17.970955+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4P5XPPGT34NNX4JBXIBDBAJTEE","json":"https://pith.science/pith/4P5XPPGT34NNX4JBXIBDBAJTEE.json","graph_json":"https://pith.science/api/pith-number/4P5XPPGT34NNX4JBXIBDBAJTEE/graph.json","events_json":"https://pith.science/api/pith-number/4P5XPPGT34NNX4JBXIBDBAJTEE/events.json","paper":"https://pith.science/paper/4P5XPPGT"},"agent_actions":{"view_html":"https://pith.science/pith/4P5XPPGT34NNX4JBXIBDBAJTEE","download_json":"https://pith.science/pith/4P5XPPGT34NNX4JBXIBDBAJTEE.json","view_paper":"https://pith.science/paper/4P5XPPGT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.01692&json=true","fetch_graph":"https://pith.science/api/pith-number/4P5XPPGT34NNX4JBXIBDBAJTEE/graph.json","fetch_events":"https://pith.science/api/pith-number/4P5XPPGT34NNX4JBXIBDBAJTEE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4P5XPPGT34NNX4JBXIBDBAJTEE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4P5XPPGT34NNX4JBXIBDBAJTEE/action/storage_attestation","attest_author":"https://pith.science/pith/4P5XPPGT34NNX4JBXIBDBAJTEE/action/author_attestation","sign_citation":"https://pith.science/pith/4P5XPPGT34NNX4JBXIBDBAJTEE/action/citation_signature","submit_replication":"https://pith.science/pith/4P5XPPGT34NNX4JBXIBDBAJTEE/action/replication_record"}},"created_at":"2026-07-05T11:14:17.970955+00:00","updated_at":"2026-07-05T11:14:17.970955+00:00"}