{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:3YPHQM4RQGIRSD3CZNCPQRLUZD","short_pith_number":"pith:3YPHQM4R","canonical_record":{"source":{"id":"2410.05193","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-07T16:50:47Z","cross_cats_sorted":[],"title_canon_sha256":"8516e853a8c5ff1fca6a8e19bb42ac9dd758bf45917fe825ea5dd7be48b1f008","abstract_canon_sha256":"b81733b78fac45e432ac91ce94436dca674898b4cb889dec077e626e6aa7b9b7"},"schema_version":"1.0"},"canonical_sha256":"de1e7833918191190f62cb44f84574c8cebf14e2500e2765e675a2f737a86401","source":{"kind":"arxiv","id":"2410.05193","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.05193","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"arxiv_version","alias_value":"2410.05193v3","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.05193","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"pith_short_12","alias_value":"3YPHQM4RQGIR","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"pith_short_16","alias_value":"3YPHQM4RQGIRSD3C","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"pith_short_8","alias_value":"3YPHQM4R","created_at":"2026-07-05T10:45:12Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:3YPHQM4RQGIRSD3CZNCPQRLUZD","target":"record","payload":{"canonical_record":{"source":{"id":"2410.05193","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-07T16:50:47Z","cross_cats_sorted":[],"title_canon_sha256":"8516e853a8c5ff1fca6a8e19bb42ac9dd758bf45917fe825ea5dd7be48b1f008","abstract_canon_sha256":"b81733b78fac45e432ac91ce94436dca674898b4cb889dec077e626e6aa7b9b7"},"schema_version":"1.0"},"canonical_sha256":"de1e7833918191190f62cb44f84574c8cebf14e2500e2765e675a2f737a86401","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:45:12.337410Z","signature_b64":"cIp4xsnO4qi48pQ/XB1OvfaArYief6dPzMeD991sXtoRMHO3fJQ/2/pC6xTamzY5Mxt7ksArsRay6+89UouTDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"de1e7833918191190f62cb44f84574c8cebf14e2500e2765e675a2f737a86401","last_reissued_at":"2026-07-05T10:45:12.336901Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:45:12.336901Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.05193","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:45:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CBmozEs1EKosNvf455Z0xqIAQ8iZX3fS8jQPUoWrtlaEcEkpnLRt2IZCYk/EMmJ5qXdxgsefYTHoV3jS8wyMDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-23T10:54:11.011594Z"},"content_sha256":"7864ecf4297047e3a4fb6e678f24231095d4b351fc4d9ba8e335f1ef21060b89","schema_version":"1.0","event_id":"sha256:7864ecf4297047e3a4fb6e678f24231095d4b351fc4d9ba8e335f1ef21060b89"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:3YPHQM4RQGIRSD3CZNCPQRLUZD","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"RevisEval: Improving LLM-as-a-Judge via Response-Adapted References","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chen Ma, Chuhan Wu, Fuyuan Lyu, Liangyou Li, Lifeng Shang, Qiyuan Zhang, Ruiming Tang, Tiezheng Yu, Xin Jiang, Yasheng Wang, Yufei Wang, Yuxin Jiang","submitted_at":"2024-10-07T16:50:47Z","abstract_excerpt":"With significant efforts in recent studies, LLM-as-a-Judge has become a cost-effective alternative to human evaluation for assessing text generation quality in a wide range of tasks. However, there still remains a reliability gap between LLM-as-a-Judge and human evaluation. One important reason is the lack of guided oracles in the evaluation process. Motivated by the role of reference pervasively used in classic text evaluation, we introduce RevisEval, a novel text generation evaluation paradigm via the response-adapted references. RevisEval is driven by the key observation that an ideal refer"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.05193","kind":"arxiv","version":3},"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/2410.05193/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":"2026-07-05T10:45:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DLZoltAtQdVrtsvwM+izEfzHCJ7427iATAn0ybpVhFehzozUIMYVByxlh7WAmoGmpCAoVrEv72GPb0JtSjX4Cw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-23T10:54:11.012120Z"},"content_sha256":"e0b3f9a12b4245ce3ceffeb29ec0504a811a9e67acefaa2153efdb8899e65472","schema_version":"1.0","event_id":"sha256:e0b3f9a12b4245ce3ceffeb29ec0504a811a9e67acefaa2153efdb8899e65472"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/3YPHQM4RQGIRSD3CZNCPQRLUZD/bundle.json","state_url":"https://pith.science/pith/3YPHQM4RQGIRSD3CZNCPQRLUZD/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/3YPHQM4RQGIRSD3CZNCPQRLUZD/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":"2026-08-23T10:54:11Z","links":{"resolver":"https://pith.science/pith/3YPHQM4RQGIRSD3CZNCPQRLUZD","bundle":"https://pith.science/pith/3YPHQM4RQGIRSD3CZNCPQRLUZD/bundle.json","state":"https://pith.science/pith/3YPHQM4RQGIRSD3CZNCPQRLUZD/state.json","well_known_bundle":"https://pith.science/.well-known/pith/3YPHQM4RQGIRSD3CZNCPQRLUZD/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:3YPHQM4RQGIRSD3CZNCPQRLUZD","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":"b81733b78fac45e432ac91ce94436dca674898b4cb889dec077e626e6aa7b9b7","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-07T16:50:47Z","title_canon_sha256":"8516e853a8c5ff1fca6a8e19bb42ac9dd758bf45917fe825ea5dd7be48b1f008"},"schema_version":"1.0","source":{"id":"2410.05193","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.05193","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"arxiv_version","alias_value":"2410.05193v3","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.05193","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"pith_short_12","alias_value":"3YPHQM4RQGIR","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"pith_short_16","alias_value":"3YPHQM4RQGIRSD3C","created_at":"2026-07-05T10:45:12Z"},{"alias_kind":"pith_short_8","alias_value":"3YPHQM4R","created_at":"2026-07-05T10:45:12Z"}],"graph_snapshots":[{"event_id":"sha256:e0b3f9a12b4245ce3ceffeb29ec0504a811a9e67acefaa2153efdb8899e65472","target":"graph","created_at":"2026-07-05T10:45:12Z","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/2410.05193/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"With significant efforts in recent studies, LLM-as-a-Judge has become a cost-effective alternative to human evaluation for assessing text generation quality in a wide range of tasks. However, there still remains a reliability gap between LLM-as-a-Judge and human evaluation. One important reason is the lack of guided oracles in the evaluation process. Motivated by the role of reference pervasively used in classic text evaluation, we introduce RevisEval, a novel text generation evaluation paradigm via the response-adapted references. RevisEval is driven by the key observation that an ideal refer","authors_text":"Chen Ma, Chuhan Wu, Fuyuan Lyu, Liangyou Li, Lifeng Shang, Qiyuan Zhang, Ruiming Tang, Tiezheng Yu, Xin Jiang, Yasheng Wang, Yufei Wang, Yuxin Jiang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-07T16:50:47Z","title":"RevisEval: Improving LLM-as-a-Judge via Response-Adapted References"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.05193","kind":"arxiv","version":3},"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":"sha256:7864ecf4297047e3a4fb6e678f24231095d4b351fc4d9ba8e335f1ef21060b89","target":"record","created_at":"2026-07-05T10:45:12Z","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":"b81733b78fac45e432ac91ce94436dca674898b4cb889dec077e626e6aa7b9b7","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-07T16:50:47Z","title_canon_sha256":"8516e853a8c5ff1fca6a8e19bb42ac9dd758bf45917fe825ea5dd7be48b1f008"},"schema_version":"1.0","source":{"id":"2410.05193","kind":"arxiv","version":3}},"canonical_sha256":"de1e7833918191190f62cb44f84574c8cebf14e2500e2765e675a2f737a86401","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"de1e7833918191190f62cb44f84574c8cebf14e2500e2765e675a2f737a86401","first_computed_at":"2026-07-05T10:45:12.336901Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:45:12.336901Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"cIp4xsnO4qi48pQ/XB1OvfaArYief6dPzMeD991sXtoRMHO3fJQ/2/pC6xTamzY5Mxt7ksArsRay6+89UouTDw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:45:12.337410Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.05193","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7864ecf4297047e3a4fb6e678f24231095d4b351fc4d9ba8e335f1ef21060b89","sha256:e0b3f9a12b4245ce3ceffeb29ec0504a811a9e67acefaa2153efdb8899e65472"],"state_sha256":"62cefb0aeff1dc15a6133fc4b7499ffc276b9210e970568ef345d76861c5d3cd"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oBJR+KctcDpeomLYC2ZqpJ/4H4+vmXdhDgV56FNUo29e2GFeQ2DgDaLj4s5rkPCOUZWmD99rtIuHnQa2SJOmDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-23T10:54:11.016427Z","bundle_sha256":"8f14232048dedf3c5813c754f7bbb275d1cc7f1da0d4028093946854cadd8bc5"}}