{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4AME2ISUSHHU4CFR6J772F4I4N","short_pith_number":"pith:4AME2ISU","schema_version":"1.0","canonical_sha256":"e0184d225491cf4e08b1f27ffd1788e3799c367d187f416820e585ce23a603f3","source":{"kind":"arxiv","id":"2606.17634","version":1},"attestation_state":"computed","paper":{"title":"Prompt Perturbation for Reliable LLM Evaluation over Comparison Graphs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dong Huang, Jianbo Sun, Pengkun Yang","submitted_at":"2026-06-16T07:44:45Z","abstract_excerpt":"Evaluating large language models (LLMs) is important for understanding their capabilities, comparing competing systems, and supporting the deployment of reliable models in practice. For open-ended tasks, pairwise evaluation has become a popular paradigm, in which two responses to the same prompt are compared and the resulting judgments are aggregated into an overall ranking. A central challenge of this paradigm is intransitivity: the induced comparison outcomes may fail to support any coherent global ranking. For example, one may observe cyclic preferences such as $A \\succ B \\succ C \\succ A$, "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.17634","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-06-16T07:44:45Z","cross_cats_sorted":[],"title_canon_sha256":"fdc639872f881e0da22985ba983b84461cb9dec3fc192cfa5d214b875a4b6314","abstract_canon_sha256":"35ee847229f8423aef375face6d67886963aa2650abd191e2a0ee6624607a965"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:10:17.780044Z","signature_b64":"0GDnerp5+uVzTXHXEFyxqGn18iIeXaIM61Ro31ssuBdAO4M/znJozeK7YwzLv58aMYX0THPkoIn0vZI8EkLzCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e0184d225491cf4e08b1f27ffd1788e3799c367d187f416820e585ce23a603f3","last_reissued_at":"2026-06-19T16:10:17.779705Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:10:17.779705Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Prompt Perturbation for Reliable LLM Evaluation over Comparison Graphs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dong Huang, Jianbo Sun, Pengkun Yang","submitted_at":"2026-06-16T07:44:45Z","abstract_excerpt":"Evaluating large language models (LLMs) is important for understanding their capabilities, comparing competing systems, and supporting the deployment of reliable models in practice. For open-ended tasks, pairwise evaluation has become a popular paradigm, in which two responses to the same prompt are compared and the resulting judgments are aggregated into an overall ranking. A central challenge of this paradigm is intransitivity: the induced comparison outcomes may fail to support any coherent global ranking. For example, one may observe cyclic preferences such as $A \\succ B \\succ C \\succ A$, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.17634","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/2606.17634/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.17634","created_at":"2026-06-19T16:10:17.779766+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.17634v1","created_at":"2026-06-19T16:10:17.779766+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.17634","created_at":"2026-06-19T16:10:17.779766+00:00"},{"alias_kind":"pith_short_12","alias_value":"4AME2ISUSHHU","created_at":"2026-06-19T16:10:17.779766+00:00"},{"alias_kind":"pith_short_16","alias_value":"4AME2ISUSHHU4CFR","created_at":"2026-06-19T16:10:17.779766+00:00"},{"alias_kind":"pith_short_8","alias_value":"4AME2ISU","created_at":"2026-06-19T16:10:17.779766+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4AME2ISUSHHU4CFR6J772F4I4N","json":"https://pith.science/pith/4AME2ISUSHHU4CFR6J772F4I4N.json","graph_json":"https://pith.science/api/pith-number/4AME2ISUSHHU4CFR6J772F4I4N/graph.json","events_json":"https://pith.science/api/pith-number/4AME2ISUSHHU4CFR6J772F4I4N/events.json","paper":"https://pith.science/paper/4AME2ISU"},"agent_actions":{"view_html":"https://pith.science/pith/4AME2ISUSHHU4CFR6J772F4I4N","download_json":"https://pith.science/pith/4AME2ISUSHHU4CFR6J772F4I4N.json","view_paper":"https://pith.science/paper/4AME2ISU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.17634&json=true","fetch_graph":"https://pith.science/api/pith-number/4AME2ISUSHHU4CFR6J772F4I4N/graph.json","fetch_events":"https://pith.science/api/pith-number/4AME2ISUSHHU4CFR6J772F4I4N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4AME2ISUSHHU4CFR6J772F4I4N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4AME2ISUSHHU4CFR6J772F4I4N/action/storage_attestation","attest_author":"https://pith.science/pith/4AME2ISUSHHU4CFR6J772F4I4N/action/author_attestation","sign_citation":"https://pith.science/pith/4AME2ISUSHHU4CFR6J772F4I4N/action/citation_signature","submit_replication":"https://pith.science/pith/4AME2ISUSHHU4CFR6J772F4I4N/action/replication_record"}},"created_at":"2026-06-19T16:10:17.779766+00:00","updated_at":"2026-06-19T16:10:17.779766+00:00"}