{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:5O4X53VFRZSJN4Z74KL5FC7T4M","short_pith_number":"pith:5O4X53VF","canonical_record":{"source":{"id":"2410.23726","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-10-31T08:26:51Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"a753439bddb5dcd458affc3d8698d3ff2d917fca397f078ce41229fc3991881c","abstract_canon_sha256":"a981823ea79d39074bff2105eca616985f68d748377a2184457f236cf9cc0895"},"schema_version":"1.0"},"canonical_sha256":"ebb97eeea58e6496f33fe297d28bf3e310c1b3a3f6af715a4c65320452645e17","source":{"kind":"arxiv","id":"2410.23726","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.23726","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"arxiv_version","alias_value":"2410.23726v1","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.23726","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"pith_short_12","alias_value":"5O4X53VFRZSJ","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"pith_short_16","alias_value":"5O4X53VFRZSJN4Z7","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"pith_short_8","alias_value":"5O4X53VF","created_at":"2026-07-05T09:29:04Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:5O4X53VFRZSJN4Z74KL5FC7T4M","target":"record","payload":{"canonical_record":{"source":{"id":"2410.23726","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-10-31T08:26:51Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"a753439bddb5dcd458affc3d8698d3ff2d917fca397f078ce41229fc3991881c","abstract_canon_sha256":"a981823ea79d39074bff2105eca616985f68d748377a2184457f236cf9cc0895"},"schema_version":"1.0"},"canonical_sha256":"ebb97eeea58e6496f33fe297d28bf3e310c1b3a3f6af715a4c65320452645e17","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:29:04.442888Z","signature_b64":"cS3NoPOum5YDsTmCR3Ksi6NHnlMJHgDuy3nbld9E9e11709HgFb8OPPujyMdPep1NJVV35rOWMqOtJbtNrUaDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ebb97eeea58e6496f33fe297d28bf3e310c1b3a3f6af715a4c65320452645e17","last_reissued_at":"2026-07-05T09:29:04.442376Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:29:04.442376Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.23726","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":"2026-07-05T09:29:04Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jv2CSY5+GCUsLZbj46sKqwh+EAppri2TfKWmqk/jMPXW4eM6794qI5OWKu9gQxWxUr600f6XxIYdt1KjbYIuCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T14:52:42.422084Z"},"content_sha256":"e580a1cd7dd1c2ab0a11127788c36b7d3b15ef2ad1e57df19faf1fe6dfb96d0d","schema_version":"1.0","event_id":"sha256:e580a1cd7dd1c2ab0a11127788c36b7d3b15ef2ad1e57df19faf1fe6dfb96d0d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:5O4X53VFRZSJN4Z74KL5FC7T4M","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Towards Reliable Alignment: Uncertainty-aware RLHF","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Aditya Gopalan, Debangshu Banerjee","submitted_at":"2024-10-31T08:26:51Z","abstract_excerpt":"Recent advances in aligning Large Language Models with human preferences have benefited from larger reward models and better preference data. However, most of these methodologies rely on the accuracy of the reward model. The reward models used in Reinforcement Learning with Human Feedback (RLHF) are typically learned from small datasets using stochastic optimization algorithms, making them prone to high variability. We illustrate the inconsistencies between reward models empirically on numerous open-source datasets.\n  We theoretically show that the fluctuation of the reward models can be detri"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.23726","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/2410.23726/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":"2026-07-05T09:29:04Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nZrRzy2uRlITRveLjZRGRLAR6RxA11CQPlYXJ3kJe9pgTs4zOD8BYxfv8AXaua9CjWdgYz9B3dF3oCH2pv3kBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T14:52:42.423041Z"},"content_sha256":"00f6d3694332537813d5e10a792fdeed83f7f2ec82bdc19b47aab275e5309c07","schema_version":"1.0","event_id":"sha256:00f6d3694332537813d5e10a792fdeed83f7f2ec82bdc19b47aab275e5309c07"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5O4X53VFRZSJN4Z74KL5FC7T4M/bundle.json","state_url":"https://pith.science/pith/5O4X53VFRZSJN4Z74KL5FC7T4M/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5O4X53VFRZSJN4Z74KL5FC7T4M/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":"2026-08-13T14:52:42Z","links":{"resolver":"https://pith.science/pith/5O4X53VFRZSJN4Z74KL5FC7T4M","bundle":"https://pith.science/pith/5O4X53VFRZSJN4Z74KL5FC7T4M/bundle.json","state":"https://pith.science/pith/5O4X53VFRZSJN4Z74KL5FC7T4M/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5O4X53VFRZSJN4Z74KL5FC7T4M/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:5O4X53VFRZSJN4Z74KL5FC7T4M","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":"a981823ea79d39074bff2105eca616985f68d748377a2184457f236cf9cc0895","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-10-31T08:26:51Z","title_canon_sha256":"a753439bddb5dcd458affc3d8698d3ff2d917fca397f078ce41229fc3991881c"},"schema_version":"1.0","source":{"id":"2410.23726","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.23726","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"arxiv_version","alias_value":"2410.23726v1","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.23726","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"pith_short_12","alias_value":"5O4X53VFRZSJ","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"pith_short_16","alias_value":"5O4X53VFRZSJN4Z7","created_at":"2026-07-05T09:29:04Z"},{"alias_kind":"pith_short_8","alias_value":"5O4X53VF","created_at":"2026-07-05T09:29:04Z"}],"graph_snapshots":[{"event_id":"sha256:00f6d3694332537813d5e10a792fdeed83f7f2ec82bdc19b47aab275e5309c07","target":"graph","created_at":"2026-07-05T09:29:04Z","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/2410.23726/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advances in aligning Large Language Models with human preferences have benefited from larger reward models and better preference data. However, most of these methodologies rely on the accuracy of the reward model. The reward models used in Reinforcement Learning with Human Feedback (RLHF) are typically learned from small datasets using stochastic optimization algorithms, making them prone to high variability. We illustrate the inconsistencies between reward models empirically on numerous open-source datasets.\n  We theoretically show that the fluctuation of the reward models can be detri","authors_text":"Aditya Gopalan, Debangshu Banerjee","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-10-31T08:26:51Z","title":"Towards Reliable Alignment: Uncertainty-aware RLHF"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.23726","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:e580a1cd7dd1c2ab0a11127788c36b7d3b15ef2ad1e57df19faf1fe6dfb96d0d","target":"record","created_at":"2026-07-05T09:29:04Z","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":"a981823ea79d39074bff2105eca616985f68d748377a2184457f236cf9cc0895","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-10-31T08:26:51Z","title_canon_sha256":"a753439bddb5dcd458affc3d8698d3ff2d917fca397f078ce41229fc3991881c"},"schema_version":"1.0","source":{"id":"2410.23726","kind":"arxiv","version":1}},"canonical_sha256":"ebb97eeea58e6496f33fe297d28bf3e310c1b3a3f6af715a4c65320452645e17","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ebb97eeea58e6496f33fe297d28bf3e310c1b3a3f6af715a4c65320452645e17","first_computed_at":"2026-07-05T09:29:04.442376Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:29:04.442376Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"cS3NoPOum5YDsTmCR3Ksi6NHnlMJHgDuy3nbld9E9e11709HgFb8OPPujyMdPep1NJVV35rOWMqOtJbtNrUaDg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:29:04.442888Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.23726","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e580a1cd7dd1c2ab0a11127788c36b7d3b15ef2ad1e57df19faf1fe6dfb96d0d","sha256:00f6d3694332537813d5e10a792fdeed83f7f2ec82bdc19b47aab275e5309c07"],"state_sha256":"744ff60bb5a6d5c2984975158ed3e2721333ee50b28911ff39eeac97ec87563e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JvdXPddGAEOtzYpZjaVztVglSmVxuyOcdw0JHZgEIx05E5USPCX4+h3c/RCLWlg2+kk+3sL7umSk4C+Neiw7Cg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-13T14:52:42.509027Z","bundle_sha256":"0b3baf665a3a0e020b4eced06120bde60c80af7404ed1f5860d2ac9e70c93701"}}