{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:6CWT6K3QKRZA7DN5XXOXYGHCBN","short_pith_number":"pith:6CWT6K3Q","canonical_record":{"source":{"id":"2604.03238","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-01-31T21:51:17Z","cross_cats_sorted":[],"title_canon_sha256":"e4a26c9ac93ebfd2ce3f2d4f0503e2bca3054eae9a56831c16cac577c9f94a3a","abstract_canon_sha256":"d2aea64851723e357251f0578e5c3cf72d321564641d89d433bab1e8065e739c"},"schema_version":"1.0"},"canonical_sha256":"f0ad3f2b7054720f8dbdbddd7c18e20b444390f70bce50612bec69aefa1999ae","source":{"kind":"arxiv","id":"2604.03238","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.03238","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"arxiv_version","alias_value":"2604.03238v2","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.03238","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"pith_short_12","alias_value":"6CWT6K3QKRZA","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"pith_short_16","alias_value":"6CWT6K3QKRZA7DN5","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"pith_short_8","alias_value":"6CWT6K3Q","created_at":"2026-06-02T01:03:46Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:6CWT6K3QKRZA7DN5XXOXYGHCBN","target":"record","payload":{"canonical_record":{"source":{"id":"2604.03238","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-01-31T21:51:17Z","cross_cats_sorted":[],"title_canon_sha256":"e4a26c9ac93ebfd2ce3f2d4f0503e2bca3054eae9a56831c16cac577c9f94a3a","abstract_canon_sha256":"d2aea64851723e357251f0578e5c3cf72d321564641d89d433bab1e8065e739c"},"schema_version":"1.0"},"canonical_sha256":"f0ad3f2b7054720f8dbdbddd7c18e20b444390f70bce50612bec69aefa1999ae","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:46.566420Z","signature_b64":"RJsL8lrx2urR40czhSxmN0sZaEt810whLJynWqSoNEoRehymuxzYi5IlVyrI6oG1d8HsxTnm8ATFsu/MAfUUDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f0ad3f2b7054720f8dbdbddd7c18e20b444390f70bce50612bec69aefa1999ae","last_reissued_at":"2026-06-02T01:03:46.565674Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:46.565674Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.03238","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":"2026-06-02T01:03:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"bC7de6UzoUH3i5W5R7o0E7jq2esVJTDFKB5ZiDdODDoLackYlV3IYakkwZveUb3R8lefYuQDVeMft6ulz4KEDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-29T19:38:34.497263Z"},"content_sha256":"1569e4f5a0d77c6038d40c0db03dc7e8f4d6fab0f593991bdafb2457196dc79b","schema_version":"1.0","event_id":"sha256:1569e4f5a0d77c6038d40c0db03dc7e8f4d6fab0f593991bdafb2457196dc79b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:6CWT6K3QKRZA7DN5XXOXYGHCBN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"RLHF May Not Reflect Genuine Preferences","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Measuring human preferences in RLHF is a social science problem because responses often reflect non-attitudes and constructed preferences rather than genuine opinions.","cross_cats":[],"primary_cat":"cs.HC","authors_text":"Bijean Ghafouri, Emilio Ferrara, Eun Cheol Choi, Priyanka Dey","submitted_at":"2026-01-31T21:51:17Z","abstract_excerpt":"Reinforcement Learning from Human Feedback (RLHF) assumes that annotation responses reflect genuine human preferences. They often do not. Behavioral scientists have documented for sixty years that people produce responses without holding genuine opinions, construct preferences on the spot from contextual cues, and interpret identical questions differently. Importantly, these failures are common for the judgments on values that matter most for AI alignment. We argue that measurement validity is logically prior to preference aggregation. Before asking how to combine annotations, the field must a"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We argue that measuring human preferences in RLHF is a social science problem. We present a taxonomy distinguishing genuine preferences from non-attitudes, constructed preferences, and measurement artifacts, along with diagnostic approaches for detecting each.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the documented phenomena of non-attitudes and constructed preferences from general behavioral science literature transfer directly and pervasively to the specific task of RLHF annotation without needing domain-specific validation studies.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"RLHF preference measurement is a social science validity problem because annotators routinely produce non-attitudes, constructed responses, and artifacts rather than stable values.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Measuring human preferences in RLHF is a social science problem because responses often reflect non-attitudes and constructed preferences rather than genuine opinions.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"7702de791ea496775b46878dc8e1e05bc2dd3d442c5a4429b28ab498911ce80c"},"source":{"id":"2604.03238","kind":"arxiv","version":2},"verdict":{"id":"93ca9d3a-67d9-4903-803d-ab50933a2575","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T08:36:38.324201Z","strongest_claim":"We argue that measuring human preferences in RLHF is a social science problem. We present a taxonomy distinguishing genuine preferences from non-attitudes, constructed preferences, and measurement artifacts, along with diagnostic approaches for detecting each.","one_line_summary":"RLHF preference measurement is a social science validity problem because annotators routinely produce non-attitudes, constructed responses, and artifacts rather than stable values.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the documented phenomena of non-attitudes and constructed preferences from general behavioral science literature transfer directly and pervasively to the specific task of RLHF annotation without needing domain-specific validation studies.","pith_extraction_headline":"Measuring human preferences in RLHF is a social science problem because responses often reflect non-attitudes and constructed preferences rather than genuine opinions."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.03238/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":2,"snapshot_sha256":"b399449970109bd8b823ecc532c2700ff7a46c3f37929ab449f7edba2c0430ca"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"93ca9d3a-67d9-4903-803d-ab50933a2575"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-02T01:03:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ELQR1W/PseU3KzRPa7umhiZSJatViobQPg8jDn1ug57p9LeKbc7al1DjdbmHy0kSRKVjCKbnrg/r9pX3EArYAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-29T19:38:34.497763Z"},"content_sha256":"c7173457f78fc1be3d4f7804c8ab566d190ef570a07f658d35795ba9f5ba727e","schema_version":"1.0","event_id":"sha256:c7173457f78fc1be3d4f7804c8ab566d190ef570a07f658d35795ba9f5ba727e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6CWT6K3QKRZA7DN5XXOXYGHCBN/bundle.json","state_url":"https://pith.science/pith/6CWT6K3QKRZA7DN5XXOXYGHCBN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6CWT6K3QKRZA7DN5XXOXYGHCBN/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":"2026-06-29T19:38:34Z","links":{"resolver":"https://pith.science/pith/6CWT6K3QKRZA7DN5XXOXYGHCBN","bundle":"https://pith.science/pith/6CWT6K3QKRZA7DN5XXOXYGHCBN/bundle.json","state":"https://pith.science/pith/6CWT6K3QKRZA7DN5XXOXYGHCBN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6CWT6K3QKRZA7DN5XXOXYGHCBN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:6CWT6K3QKRZA7DN5XXOXYGHCBN","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":"d2aea64851723e357251f0578e5c3cf72d321564641d89d433bab1e8065e739c","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-01-31T21:51:17Z","title_canon_sha256":"e4a26c9ac93ebfd2ce3f2d4f0503e2bca3054eae9a56831c16cac577c9f94a3a"},"schema_version":"1.0","source":{"id":"2604.03238","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.03238","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"arxiv_version","alias_value":"2604.03238v2","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.03238","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"pith_short_12","alias_value":"6CWT6K3QKRZA","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"pith_short_16","alias_value":"6CWT6K3QKRZA7DN5","created_at":"2026-06-02T01:03:46Z"},{"alias_kind":"pith_short_8","alias_value":"6CWT6K3Q","created_at":"2026-06-02T01:03:46Z"}],"graph_snapshots":[{"event_id":"sha256:c7173457f78fc1be3d4f7804c8ab566d190ef570a07f658d35795ba9f5ba727e","target":"graph","created_at":"2026-06-02T01:03:46Z","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":"We argue that measuring human preferences in RLHF is a social science problem. We present a taxonomy distinguishing genuine preferences from non-attitudes, constructed preferences, and measurement artifacts, along with diagnostic approaches for detecting each."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the documented phenomena of non-attitudes and constructed preferences from general behavioral science literature transfer directly and pervasively to the specific task of RLHF annotation without needing domain-specific validation studies."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"RLHF preference measurement is a social science validity problem because annotators routinely produce non-attitudes, constructed responses, and artifacts rather than stable values."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Measuring human preferences in RLHF is a social science problem because responses often reflect non-attitudes and constructed preferences rather than genuine opinions."}],"snapshot_sha256":"7702de791ea496775b46878dc8e1e05bc2dd3d442c5a4429b28ab498911ce80c"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"b399449970109bd8b823ecc532c2700ff7a46c3f37929ab449f7edba2c0430ca"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.03238/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement Learning from Human Feedback (RLHF) assumes that annotation responses reflect genuine human preferences. They often do not. Behavioral scientists have documented for sixty years that people produce responses without holding genuine opinions, construct preferences on the spot from contextual cues, and interpret identical questions differently. Importantly, these failures are common for the judgments on values that matter most for AI alignment. We argue that measurement validity is logically prior to preference aggregation. Before asking how to combine annotations, the field must a","authors_text":"Bijean Ghafouri, Emilio Ferrara, Eun Cheol Choi, Priyanka Dey","cross_cats":[],"headline":"Measuring human preferences in RLHF is a social science problem because responses often reflect non-attitudes and constructed preferences rather than genuine opinions.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-01-31T21:51:17Z","title":"RLHF May Not Reflect Genuine Preferences"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.03238","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-16T08:36:38.324201Z","id":"93ca9d3a-67d9-4903-803d-ab50933a2575","model_set":{"reader":"grok-4.3"},"one_line_summary":"RLHF preference measurement is a social science validity problem because annotators routinely produce non-attitudes, constructed responses, and artifacts rather than stable values.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Measuring human preferences in RLHF is a social science problem because responses often reflect non-attitudes and constructed preferences rather than genuine opinions.","strongest_claim":"We argue that measuring human preferences in RLHF is a social science problem. We present a taxonomy distinguishing genuine preferences from non-attitudes, constructed preferences, and measurement artifacts, along with diagnostic approaches for detecting each.","weakest_assumption":"That the documented phenomena of non-attitudes and constructed preferences from general behavioral science literature transfer directly and pervasively to the specific task of RLHF annotation without needing domain-specific validation studies."}},"verdict_id":"93ca9d3a-67d9-4903-803d-ab50933a2575"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:1569e4f5a0d77c6038d40c0db03dc7e8f4d6fab0f593991bdafb2457196dc79b","target":"record","created_at":"2026-06-02T01:03:46Z","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":"d2aea64851723e357251f0578e5c3cf72d321564641d89d433bab1e8065e739c","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-01-31T21:51:17Z","title_canon_sha256":"e4a26c9ac93ebfd2ce3f2d4f0503e2bca3054eae9a56831c16cac577c9f94a3a"},"schema_version":"1.0","source":{"id":"2604.03238","kind":"arxiv","version":2}},"canonical_sha256":"f0ad3f2b7054720f8dbdbddd7c18e20b444390f70bce50612bec69aefa1999ae","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f0ad3f2b7054720f8dbdbddd7c18e20b444390f70bce50612bec69aefa1999ae","first_computed_at":"2026-06-02T01:03:46.565674Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T01:03:46.565674Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"RJsL8lrx2urR40czhSxmN0sZaEt810whLJynWqSoNEoRehymuxzYi5IlVyrI6oG1d8HsxTnm8ATFsu/MAfUUDw==","signature_status":"signed_v1","signed_at":"2026-06-02T01:03:46.566420Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.03238","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1569e4f5a0d77c6038d40c0db03dc7e8f4d6fab0f593991bdafb2457196dc79b","sha256:c7173457f78fc1be3d4f7804c8ab566d190ef570a07f658d35795ba9f5ba727e"],"state_sha256":"f3e7b9c030c11287d2aac1f7cd543951462c43fc6daa4acedecdcb6ce6e63dfe"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FpqvzFEptIRkkzD7NGemshdCrJsPkN4e33IfKlGTyUEYCZsScCCAS1FCwSku4W2CeXD8pI2hm+wlR+elYu6rBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-29T19:38:34.499907Z","bundle_sha256":"04e5148c0616a571f56a1ea83b12372c5a75e97e9092621d35042a4e83dea8c8"}}