{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:6EDBOHMWVUJ2TAGDYLQW4FDXVE","short_pith_number":"pith:6EDBOHMW","schema_version":"1.0","canonical_sha256":"f106171d96ad13a980c3c2e16e1477a93764e67c49204c11e887f60a79668a9b","source":{"kind":"arxiv","id":"2503.00137","version":1},"attestation_state":"computed","paper":{"title":"SCORE: Systematic COnsistency and Robustness Evaluation for Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Evelina Bakhturina, Grigor Nalbandyan, Rima Shahbazyan","submitted_at":"2025-02-28T19:27:29Z","abstract_excerpt":"Typical evaluations of Large Language Models (LLMs) report a single metric per dataset, often representing the model's best-case performance under carefully selected settings. Unfortunately, this approach overlooks model robustness and reliability in real-world applications. For instance, simple paraphrasing of prompts on the MMLU-Pro dataset causes accuracy fluctuations of up to 10\\%, while reordering answer choices in the AGIEval dataset results in accuracy differences of up to 6.1\\%. While some studies discuss issues with LLM robustness, there is no unified or centralized framework for eval"},"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":"2503.00137","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-28T19:27:29Z","cross_cats_sorted":[],"title_canon_sha256":"bb81b753f87ea94b9be74453699dfc1f994139dcb9f2782bdc0b00b544b6a270","abstract_canon_sha256":"b6e42ecedc79fe89139694d2817d49df4753043070909b60fc493e46d5ae93e2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:22:17.863896Z","signature_b64":"vDl2Yii7hSt9M3mgHTzgPopiUQBbipT9SnddH3fu3bbfOjA4Ydddkdl35lZUeyfBmvD2WKm/+mLwcLQ2w+J3Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f106171d96ad13a980c3c2e16e1477a93764e67c49204c11e887f60a79668a9b","last_reissued_at":"2026-07-05T10:22:17.863279Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:22:17.863279Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SCORE: Systematic COnsistency and Robustness Evaluation for Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Evelina Bakhturina, Grigor Nalbandyan, Rima Shahbazyan","submitted_at":"2025-02-28T19:27:29Z","abstract_excerpt":"Typical evaluations of Large Language Models (LLMs) report a single metric per dataset, often representing the model's best-case performance under carefully selected settings. Unfortunately, this approach overlooks model robustness and reliability in real-world applications. For instance, simple paraphrasing of prompts on the MMLU-Pro dataset causes accuracy fluctuations of up to 10\\%, while reordering answer choices in the AGIEval dataset results in accuracy differences of up to 6.1\\%. While some studies discuss issues with LLM robustness, there is no unified or centralized framework for eval"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.00137","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/2503.00137/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":"2503.00137","created_at":"2026-07-05T10:22:17.863378+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.00137v1","created_at":"2026-07-05T10:22:17.863378+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.00137","created_at":"2026-07-05T10:22:17.863378+00:00"},{"alias_kind":"pith_short_12","alias_value":"6EDBOHMWVUJ2","created_at":"2026-07-05T10:22:17.863378+00:00"},{"alias_kind":"pith_short_16","alias_value":"6EDBOHMWVUJ2TAGD","created_at":"2026-07-05T10:22:17.863378+00:00"},{"alias_kind":"pith_short_8","alias_value":"6EDBOHMW","created_at":"2026-07-05T10:22:17.863378+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.27209","citing_title":"Learning to Act under Noise: Enhancing Agent Robustness via Noisy Environments","ref_index":83,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6EDBOHMWVUJ2TAGDYLQW4FDXVE","json":"https://pith.science/pith/6EDBOHMWVUJ2TAGDYLQW4FDXVE.json","graph_json":"https://pith.science/api/pith-number/6EDBOHMWVUJ2TAGDYLQW4FDXVE/graph.json","events_json":"https://pith.science/api/pith-number/6EDBOHMWVUJ2TAGDYLQW4FDXVE/events.json","paper":"https://pith.science/paper/6EDBOHMW"},"agent_actions":{"view_html":"https://pith.science/pith/6EDBOHMWVUJ2TAGDYLQW4FDXVE","download_json":"https://pith.science/pith/6EDBOHMWVUJ2TAGDYLQW4FDXVE.json","view_paper":"https://pith.science/paper/6EDBOHMW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.00137&json=true","fetch_graph":"https://pith.science/api/pith-number/6EDBOHMWVUJ2TAGDYLQW4FDXVE/graph.json","fetch_events":"https://pith.science/api/pith-number/6EDBOHMWVUJ2TAGDYLQW4FDXVE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6EDBOHMWVUJ2TAGDYLQW4FDXVE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6EDBOHMWVUJ2TAGDYLQW4FDXVE/action/storage_attestation","attest_author":"https://pith.science/pith/6EDBOHMWVUJ2TAGDYLQW4FDXVE/action/author_attestation","sign_citation":"https://pith.science/pith/6EDBOHMWVUJ2TAGDYLQW4FDXVE/action/citation_signature","submit_replication":"https://pith.science/pith/6EDBOHMWVUJ2TAGDYLQW4FDXVE/action/replication_record"}},"created_at":"2026-07-05T10:22:17.863378+00:00","updated_at":"2026-07-05T10:22:17.863378+00:00"}