{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:BLYUXNVX474BMZ62ON7YJHJBUG","short_pith_number":"pith:BLYUXNVX","canonical_record":{"source":{"id":"2410.09724","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T04:48:40Z","cross_cats_sorted":[],"title_canon_sha256":"351c4120b8f965991962fee16df2551675e83b569042e6856f51f7c3a79efb8f","abstract_canon_sha256":"7698befe89f2f7b271040ae8059f6159a5dc26aa6cc2616b27c86f28658f99e8"},"schema_version":"1.0"},"canonical_sha256":"0af14bb6b7e7f81667da737f849d21a1b6367c032ebfb69bb4d929095e1b2208","source":{"kind":"arxiv","id":"2410.09724","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.09724","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"arxiv_version","alias_value":"2410.09724v2","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.09724","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"pith_short_12","alias_value":"BLYUXNVX474B","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"pith_short_16","alias_value":"BLYUXNVX474BMZ62","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"pith_short_8","alias_value":"BLYUXNVX","created_at":"2026-07-05T10:22:07Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:BLYUXNVX474BMZ62ON7YJHJBUG","target":"record","payload":{"canonical_record":{"source":{"id":"2410.09724","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T04:48:40Z","cross_cats_sorted":[],"title_canon_sha256":"351c4120b8f965991962fee16df2551675e83b569042e6856f51f7c3a79efb8f","abstract_canon_sha256":"7698befe89f2f7b271040ae8059f6159a5dc26aa6cc2616b27c86f28658f99e8"},"schema_version":"1.0"},"canonical_sha256":"0af14bb6b7e7f81667da737f849d21a1b6367c032ebfb69bb4d929095e1b2208","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:22:07.485938Z","signature_b64":"9REK7VoE5gz7NMIPRXVIF3Z3pguK2+MU/h71OjmkiS2HmgSIWoJCEs2Tvf1eirOk12Qh6A7+mHO+yUtwCFG3DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0af14bb6b7e7f81667da737f849d21a1b6367c032ebfb69bb4d929095e1b2208","last_reissued_at":"2026-07-05T10:22:07.485374Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:22:07.485374Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.09724","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":"2026-07-05T10:22:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZpicDbnrdTIOnEdavl52JzGiPe/rTnrzxllzK2uKNB//6Il9ReHmUqKvj+vksTAeXZ4dPeuZqlNP88Q5gmq0AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T07:53:28.653578Z"},"content_sha256":"8c8b9aeb37001141194a839ab115ee09002227c78022b5805ae885bd0ddeb9cc","schema_version":"1.0","event_id":"sha256:8c8b9aeb37001141194a839ab115ee09002227c78022b5805ae885bd0ddeb9cc"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:BLYUXNVX474BMZ62ON7YJHJBUG","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Taming Overconfidence in LLMs: Reward Calibration in RLHF","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Banghua Zhu, Chengsong Huang, Jiaxin Huang, Jixuan Leng","submitted_at":"2024-10-13T04:48:40Z","abstract_excerpt":"Language model calibration refers to the alignment between the confidence of the model and the actual performance of its responses. While previous studies point out the overconfidence phenomenon in Large Language Models (LLMs) and show that LLMs trained with Reinforcement Learning from Human Feedback (RLHF) are overconfident with a more sharpened output probability, in this study, we reveal that RLHF tends to lead models to express verbalized overconfidence in their own responses. We investigate the underlying cause of this overconfidence and demonstrate that reward models used for Proximal Po"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.09724","kind":"arxiv","version":2},"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/2410.09724/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":"2026-07-05T10:22:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Cl9RAvgN7Eybf6Vv8XsiQIFGNiwmYQBcNLaeB+eTEbtvP/bMDM9uze3XReQqp5Ij13yplMcZ+yMshEdLZ2vsDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T07:53:28.654142Z"},"content_sha256":"d44df6438bec0543b1f0733c7449f33c4d5ebc31b170ee63ebb013e3608bfe95","schema_version":"1.0","event_id":"sha256:d44df6438bec0543b1f0733c7449f33c4d5ebc31b170ee63ebb013e3608bfe95"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BLYUXNVX474BMZ62ON7YJHJBUG/bundle.json","state_url":"https://pith.science/pith/BLYUXNVX474BMZ62ON7YJHJBUG/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BLYUXNVX474BMZ62ON7YJHJBUG/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":"2026-08-09T07:53:28Z","links":{"resolver":"https://pith.science/pith/BLYUXNVX474BMZ62ON7YJHJBUG","bundle":"https://pith.science/pith/BLYUXNVX474BMZ62ON7YJHJBUG/bundle.json","state":"https://pith.science/pith/BLYUXNVX474BMZ62ON7YJHJBUG/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BLYUXNVX474BMZ62ON7YJHJBUG/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:BLYUXNVX474BMZ62ON7YJHJBUG","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":"7698befe89f2f7b271040ae8059f6159a5dc26aa6cc2616b27c86f28658f99e8","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T04:48:40Z","title_canon_sha256":"351c4120b8f965991962fee16df2551675e83b569042e6856f51f7c3a79efb8f"},"schema_version":"1.0","source":{"id":"2410.09724","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.09724","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"arxiv_version","alias_value":"2410.09724v2","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.09724","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"pith_short_12","alias_value":"BLYUXNVX474B","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"pith_short_16","alias_value":"BLYUXNVX474BMZ62","created_at":"2026-07-05T10:22:07Z"},{"alias_kind":"pith_short_8","alias_value":"BLYUXNVX","created_at":"2026-07-05T10:22:07Z"}],"graph_snapshots":[{"event_id":"sha256:d44df6438bec0543b1f0733c7449f33c4d5ebc31b170ee63ebb013e3608bfe95","target":"graph","created_at":"2026-07-05T10:22:07Z","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/2410.09724/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Language model calibration refers to the alignment between the confidence of the model and the actual performance of its responses. While previous studies point out the overconfidence phenomenon in Large Language Models (LLMs) and show that LLMs trained with Reinforcement Learning from Human Feedback (RLHF) are overconfident with a more sharpened output probability, in this study, we reveal that RLHF tends to lead models to express verbalized overconfidence in their own responses. We investigate the underlying cause of this overconfidence and demonstrate that reward models used for Proximal Po","authors_text":"Banghua Zhu, Chengsong Huang, Jiaxin Huang, Jixuan Leng","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T04:48:40Z","title":"Taming Overconfidence in LLMs: Reward Calibration in RLHF"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.09724","kind":"arxiv","version":2},"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":"sha256:8c8b9aeb37001141194a839ab115ee09002227c78022b5805ae885bd0ddeb9cc","target":"record","created_at":"2026-07-05T10:22:07Z","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":"7698befe89f2f7b271040ae8059f6159a5dc26aa6cc2616b27c86f28658f99e8","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T04:48:40Z","title_canon_sha256":"351c4120b8f965991962fee16df2551675e83b569042e6856f51f7c3a79efb8f"},"schema_version":"1.0","source":{"id":"2410.09724","kind":"arxiv","version":2}},"canonical_sha256":"0af14bb6b7e7f81667da737f849d21a1b6367c032ebfb69bb4d929095e1b2208","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0af14bb6b7e7f81667da737f849d21a1b6367c032ebfb69bb4d929095e1b2208","first_computed_at":"2026-07-05T10:22:07.485374Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:22:07.485374Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"9REK7VoE5gz7NMIPRXVIF3Z3pguK2+MU/h71OjmkiS2HmgSIWoJCEs2Tvf1eirOk12Qh6A7+mHO+yUtwCFG3DQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:22:07.485938Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.09724","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8c8b9aeb37001141194a839ab115ee09002227c78022b5805ae885bd0ddeb9cc","sha256:d44df6438bec0543b1f0733c7449f33c4d5ebc31b170ee63ebb013e3608bfe95"],"state_sha256":"a15d83338e28ce91725c440ae4dc42a96cf979fd92ea3ab95e2489c6e76dd768"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4BlQttZInxrbLmWWtxgXxOG5KZJukBpuz8TaxjnyKb8u/WsZrrK19o3GmCXcctWu1daVw747i5rQVJcsNFY2BQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T07:53:28.660044Z","bundle_sha256":"73557e2f5b32a42d9279440d7ebad2f7e710d07a2775ec746013026d29868d3e"}}