{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:73IB5S7GCCAUSSMRV2NEG6A7TR","short_pith_number":"pith:73IB5S7G","canonical_record":{"source":{"id":"2402.10669","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-16T13:21:06Z","cross_cats_sorted":[],"title_canon_sha256":"c9a18699c90237d94ba48b9d55eb4552ef32c94dd12b9203626d671b79b00ab9","abstract_canon_sha256":"648b583f26c56543c2c5f69dad859dbd122846823dfbc5b4c73449019cf91e5e"},"schema_version":"1.0"},"canonical_sha256":"fed01ecbe61081494991ae9a43781f9c4a49ab3052494b6d56d31effe0e5cdd1","source":{"kind":"arxiv","id":"2402.10669","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.10669","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"arxiv_version","alias_value":"2402.10669v5","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.10669","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"pith_short_12","alias_value":"73IB5S7GCCAU","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"pith_short_16","alias_value":"73IB5S7GCCAUSSMR","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"pith_short_8","alias_value":"73IB5S7G","created_at":"2026-07-05T09:12:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:73IB5S7GCCAUSSMRV2NEG6A7TR","target":"record","payload":{"canonical_record":{"source":{"id":"2402.10669","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-16T13:21:06Z","cross_cats_sorted":[],"title_canon_sha256":"c9a18699c90237d94ba48b9d55eb4552ef32c94dd12b9203626d671b79b00ab9","abstract_canon_sha256":"648b583f26c56543c2c5f69dad859dbd122846823dfbc5b4c73449019cf91e5e"},"schema_version":"1.0"},"canonical_sha256":"fed01ecbe61081494991ae9a43781f9c4a49ab3052494b6d56d31effe0e5cdd1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:12:02.655717Z","signature_b64":"qyHVntHeO1eNw108IetlJl22I0qunaIkA3cNzHr4JBMj6sGd6RHh0WL/xK/aDLxHvmaWwz0mZU/U2i/697kPDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fed01ecbe61081494991ae9a43781f9c4a49ab3052494b6d56d31effe0e5cdd1","last_reissued_at":"2026-07-05T09:12:02.655309Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:12:02.655309Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.10669","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:12:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"t4r/xVZV+JEcz8dWnsO6ph9YprdcRA5MIRkDJ9h+YX7SU0/v4XYze/AXLeJO+vOWewgwhl7b0d1eVxKG+LaqBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T04:08:03.455134Z"},"content_sha256":"b6c799fd3f2ff08f13bc9868623d5beab364ec266eb18cda4d05fab8777ab608","schema_version":"1.0","event_id":"sha256:b6c799fd3f2ff08f13bc9868623d5beab364ec266eb18cda4d05fab8777ab608"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:73IB5S7GCCAUSSMRV2NEG6A7TR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Humans or LLMs as the Judge? A Study on Judgement Biases","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Benyou Wang, Feng Jiang, Guiming Hardy Chen, Shunian Chen, Ziche Liu","submitted_at":"2024-02-16T13:21:06Z","abstract_excerpt":"Adopting human and large language models (LLM) as judges (a.k.a human- and LLM-as-a-judge) for evaluating the performance of LLMs has recently gained attention. Nonetheless, this approach concurrently introduces potential biases from human and LLMs, questioning the reliability of the evaluation results. In this paper, we propose a novel framework that is free from referencing groundtruth annotations for investigating Misinformation Oversight Bias, Gender Bias, Authority Bias and Beauty Bias on LLM and human judges. We curate a dataset referring to the revised Bloom's Taxonomy and conduct thous"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.10669","kind":"arxiv","version":5},"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/2402.10669/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":"2026-07-05T09:12:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1RDSTwnhwbrfFgwmGdVNs0gG5i8h/+mdbVxNHzSXIe5YIMEo5xuuwciG9CYaDcWHp/tICKC+XJUZS/YG/0RACg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T04:08:03.455646Z"},"content_sha256":"dfa80285042e87357710f92c21b10b893001d21fc91e3af97c2bd413dd9efc3d","schema_version":"1.0","event_id":"sha256:dfa80285042e87357710f92c21b10b893001d21fc91e3af97c2bd413dd9efc3d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/73IB5S7GCCAUSSMRV2NEG6A7TR/bundle.json","state_url":"https://pith.science/pith/73IB5S7GCCAUSSMRV2NEG6A7TR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/73IB5S7GCCAUSSMRV2NEG6A7TR/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":"2026-08-08T04:08:03Z","links":{"resolver":"https://pith.science/pith/73IB5S7GCCAUSSMRV2NEG6A7TR","bundle":"https://pith.science/pith/73IB5S7GCCAUSSMRV2NEG6A7TR/bundle.json","state":"https://pith.science/pith/73IB5S7GCCAUSSMRV2NEG6A7TR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/73IB5S7GCCAUSSMRV2NEG6A7TR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:73IB5S7GCCAUSSMRV2NEG6A7TR","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":"648b583f26c56543c2c5f69dad859dbd122846823dfbc5b4c73449019cf91e5e","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-16T13:21:06Z","title_canon_sha256":"c9a18699c90237d94ba48b9d55eb4552ef32c94dd12b9203626d671b79b00ab9"},"schema_version":"1.0","source":{"id":"2402.10669","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.10669","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"arxiv_version","alias_value":"2402.10669v5","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.10669","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"pith_short_12","alias_value":"73IB5S7GCCAU","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"pith_short_16","alias_value":"73IB5S7GCCAUSSMR","created_at":"2026-07-05T09:12:02Z"},{"alias_kind":"pith_short_8","alias_value":"73IB5S7G","created_at":"2026-07-05T09:12:02Z"}],"graph_snapshots":[{"event_id":"sha256:dfa80285042e87357710f92c21b10b893001d21fc91e3af97c2bd413dd9efc3d","target":"graph","created_at":"2026-07-05T09:12:02Z","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/2402.10669/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Adopting human and large language models (LLM) as judges (a.k.a human- and LLM-as-a-judge) for evaluating the performance of LLMs has recently gained attention. Nonetheless, this approach concurrently introduces potential biases from human and LLMs, questioning the reliability of the evaluation results. In this paper, we propose a novel framework that is free from referencing groundtruth annotations for investigating Misinformation Oversight Bias, Gender Bias, Authority Bias and Beauty Bias on LLM and human judges. We curate a dataset referring to the revised Bloom's Taxonomy and conduct thous","authors_text":"Benyou Wang, Feng Jiang, Guiming Hardy Chen, Shunian Chen, Ziche Liu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-16T13:21:06Z","title":"Humans or LLMs as the Judge? A Study on Judgement Biases"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.10669","kind":"arxiv","version":5},"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:b6c799fd3f2ff08f13bc9868623d5beab364ec266eb18cda4d05fab8777ab608","target":"record","created_at":"2026-07-05T09:12:02Z","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":"648b583f26c56543c2c5f69dad859dbd122846823dfbc5b4c73449019cf91e5e","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-16T13:21:06Z","title_canon_sha256":"c9a18699c90237d94ba48b9d55eb4552ef32c94dd12b9203626d671b79b00ab9"},"schema_version":"1.0","source":{"id":"2402.10669","kind":"arxiv","version":5}},"canonical_sha256":"fed01ecbe61081494991ae9a43781f9c4a49ab3052494b6d56d31effe0e5cdd1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"fed01ecbe61081494991ae9a43781f9c4a49ab3052494b6d56d31effe0e5cdd1","first_computed_at":"2026-07-05T09:12:02.655309Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:12:02.655309Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"qyHVntHeO1eNw108IetlJl22I0qunaIkA3cNzHr4JBMj6sGd6RHh0WL/xK/aDLxHvmaWwz0mZU/U2i/697kPDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T09:12:02.655717Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.10669","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b6c799fd3f2ff08f13bc9868623d5beab364ec266eb18cda4d05fab8777ab608","sha256:dfa80285042e87357710f92c21b10b893001d21fc91e3af97c2bd413dd9efc3d"],"state_sha256":"60d314b1068446a2252108a18ef806b8ec99ed60e5a5035f9ded602690d90eac"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9H5HuDua9CNEmF7sIcjRs013xYBF5MaBzEDxaNXWbDx0HgNaWNS32lP/iFcHDPBR+K2xui+o/nZjM0YCZGpCBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T04:08:03.460564Z","bundle_sha256":"caa8ffc68ff63ee23ab694c93574dbc445e8c775f4d88594be4dac88e7555206"}}