{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:WPAPRZMGFB7KHRAUZUI7CNXMX3","short_pith_number":"pith:WPAPRZMG","canonical_record":{"source":{"id":"2404.08008","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-10T01:26:24Z","cross_cats_sorted":["cs.CL","cs.HC"],"title_canon_sha256":"e220484dbcbfedd224a709f17a09e0eef70df6c4b00ffffb0ed325e789abd209","abstract_canon_sha256":"0b4b8d811d1cef860ff21daced64be78263cd2af1aa7100e1f559827877d8f34"},"schema_version":"1.0"},"canonical_sha256":"b3c0f8e586287ea3c414cd11f136ecbedcb3d236a53568dbec1e513e4fc227cf","source":{"kind":"arxiv","id":"2404.08008","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.08008","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"arxiv_version","alias_value":"2404.08008v2","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.08008","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"pith_short_12","alias_value":"WPAPRZMGFB7K","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"pith_short_16","alias_value":"WPAPRZMGFB7KHRAU","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"pith_short_8","alias_value":"WPAPRZMG","created_at":"2026-07-05T11:11:31Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:WPAPRZMGFB7KHRAUZUI7CNXMX3","target":"record","payload":{"canonical_record":{"source":{"id":"2404.08008","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-10T01:26:24Z","cross_cats_sorted":["cs.CL","cs.HC"],"title_canon_sha256":"e220484dbcbfedd224a709f17a09e0eef70df6c4b00ffffb0ed325e789abd209","abstract_canon_sha256":"0b4b8d811d1cef860ff21daced64be78263cd2af1aa7100e1f559827877d8f34"},"schema_version":"1.0"},"canonical_sha256":"b3c0f8e586287ea3c414cd11f136ecbedcb3d236a53568dbec1e513e4fc227cf","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:11:31.748102Z","signature_b64":"DyA6jwxNZfYH2sjrAO47h89Wpy4BmBGc0kRsnZSaQw4bJhHveaa/PuyINLGkIRkDii/mX0iTdnah9sKiOUIJAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b3c0f8e586287ea3c414cd11f136ecbedcb3d236a53568dbec1e513e4fc227cf","last_reissued_at":"2026-07-05T11:11:31.747518Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:11:31.747518Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.08008","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":"2026-07-05T11:11:31Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mDBDkpN/riaZ218lEUn5z7qZ/Guim4oxQw7xvMRjn0TvOXMUsOvdK2aEMHsNp/U4zbPzqMCiWlq5wEu51E8+AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T20:12:28.865203Z"},"content_sha256":"837180dc0e9f3b6036335c53351437f0ab30f541a876ecc8b17af815ac31d30e","schema_version":"1.0","event_id":"sha256:837180dc0e9f3b6036335c53351437f0ab30f541a876ecc8b17af815ac31d30e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:WPAPRZMGFB7KHRAUZUI7CNXMX3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Sample-Efficient Human Evaluation of Large Language Models via Maximum Discrepancy Competition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.HC"],"primary_cat":"cs.LG","authors_text":"Ge Sun, Guozhou Zheng, Hongzhi Tan, Huajun Chen, Kede Ma, Kehua Feng, Keyan Ding, Qiang Zhang, Shuangquan Guo, Yuzhou Cheng, Zhihua Wang","submitted_at":"2024-04-10T01:26:24Z","abstract_excerpt":"Reliable evaluation of large language models (LLMs) is impeded by two key challenges: objective metrics often fail to reflect human perception of natural language, and exhaustive human labeling is prohibitively expensive. Here, we propose a sample-efficient human evaluation method for LLMs based on the principle of MAximum Discrepancy (MAD) Competition. Our method automatically and adaptively selects a compact set of input instructions that maximize semantic discrepancy between pairs of LLM responses. Human evaluators then perform three-alternative forced choices on these paired responses, whi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.08008","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/2404.08008/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":"2026-07-05T11:11:31Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pArTa/HVF6uQbyK8BkUNmU70TqUToDlQYmJ3M5VE0K/p3cbizl+vvc4w7hzgaDtButMczxqcfEUb9YJe+WLNDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T20:12:28.865734Z"},"content_sha256":"ad99e742319a104b78b895b6ad3a4b8ef31ec2fd433c45afa830307cea1d6b4b","schema_version":"1.0","event_id":"sha256:ad99e742319a104b78b895b6ad3a4b8ef31ec2fd433c45afa830307cea1d6b4b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/WPAPRZMGFB7KHRAUZUI7CNXMX3/bundle.json","state_url":"https://pith.science/pith/WPAPRZMGFB7KHRAUZUI7CNXMX3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/WPAPRZMGFB7KHRAUZUI7CNXMX3/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":"2026-08-10T20:12:28Z","links":{"resolver":"https://pith.science/pith/WPAPRZMGFB7KHRAUZUI7CNXMX3","bundle":"https://pith.science/pith/WPAPRZMGFB7KHRAUZUI7CNXMX3/bundle.json","state":"https://pith.science/pith/WPAPRZMGFB7KHRAUZUI7CNXMX3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/WPAPRZMGFB7KHRAUZUI7CNXMX3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:WPAPRZMGFB7KHRAUZUI7CNXMX3","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":"0b4b8d811d1cef860ff21daced64be78263cd2af1aa7100e1f559827877d8f34","cross_cats_sorted":["cs.CL","cs.HC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-10T01:26:24Z","title_canon_sha256":"e220484dbcbfedd224a709f17a09e0eef70df6c4b00ffffb0ed325e789abd209"},"schema_version":"1.0","source":{"id":"2404.08008","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.08008","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"arxiv_version","alias_value":"2404.08008v2","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.08008","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"pith_short_12","alias_value":"WPAPRZMGFB7K","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"pith_short_16","alias_value":"WPAPRZMGFB7KHRAU","created_at":"2026-07-05T11:11:31Z"},{"alias_kind":"pith_short_8","alias_value":"WPAPRZMG","created_at":"2026-07-05T11:11:31Z"}],"graph_snapshots":[{"event_id":"sha256:ad99e742319a104b78b895b6ad3a4b8ef31ec2fd433c45afa830307cea1d6b4b","target":"graph","created_at":"2026-07-05T11:11:31Z","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/2404.08008/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reliable evaluation of large language models (LLMs) is impeded by two key challenges: objective metrics often fail to reflect human perception of natural language, and exhaustive human labeling is prohibitively expensive. Here, we propose a sample-efficient human evaluation method for LLMs based on the principle of MAximum Discrepancy (MAD) Competition. Our method automatically and adaptively selects a compact set of input instructions that maximize semantic discrepancy between pairs of LLM responses. Human evaluators then perform three-alternative forced choices on these paired responses, whi","authors_text":"Ge Sun, Guozhou Zheng, Hongzhi Tan, Huajun Chen, Kede Ma, Kehua Feng, Keyan Ding, Qiang Zhang, Shuangquan Guo, Yuzhou Cheng, Zhihua Wang","cross_cats":["cs.CL","cs.HC"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-10T01:26:24Z","title":"Sample-Efficient Human Evaluation of Large Language Models via Maximum Discrepancy Competition"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.08008","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":"sha256:837180dc0e9f3b6036335c53351437f0ab30f541a876ecc8b17af815ac31d30e","target":"record","created_at":"2026-07-05T11:11:31Z","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":"0b4b8d811d1cef860ff21daced64be78263cd2af1aa7100e1f559827877d8f34","cross_cats_sorted":["cs.CL","cs.HC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-10T01:26:24Z","title_canon_sha256":"e220484dbcbfedd224a709f17a09e0eef70df6c4b00ffffb0ed325e789abd209"},"schema_version":"1.0","source":{"id":"2404.08008","kind":"arxiv","version":2}},"canonical_sha256":"b3c0f8e586287ea3c414cd11f136ecbedcb3d236a53568dbec1e513e4fc227cf","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b3c0f8e586287ea3c414cd11f136ecbedcb3d236a53568dbec1e513e4fc227cf","first_computed_at":"2026-07-05T11:11:31.747518Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:11:31.747518Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"DyA6jwxNZfYH2sjrAO47h89Wpy4BmBGc0kRsnZSaQw4bJhHveaa/PuyINLGkIRkDii/mX0iTdnah9sKiOUIJAg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:11:31.748102Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.08008","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:837180dc0e9f3b6036335c53351437f0ab30f541a876ecc8b17af815ac31d30e","sha256:ad99e742319a104b78b895b6ad3a4b8ef31ec2fd433c45afa830307cea1d6b4b"],"state_sha256":"8434c71a74f8cb9da9ba882043f45775a690dec9d35ae2f8f12894ea9fea2056"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"v4xPlUakoqrkcainoZdf8VQLG4cvp30zvF2gOjUQpW9MMFTiWaAtCcBGejQsqQVMffQdPjA/HTgchBXA11cBCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T20:12:28.869066Z","bundle_sha256":"6435dff0d5e6d884137d8c6a1bbe43b5e2f84ac5e57b8b6a44817ae769f874a6"}}