{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:NALMLT5X6QTYLD6CT2T7NVLTNT","short_pith_number":"pith:NALMLT5X","schema_version":"1.0","canonical_sha256":"6816c5cfb7f427858fc29ea7f6d5736cc7423c3a15573ea6c215fcf14e34ed0e","source":{"kind":"arxiv","id":"2110.08420","version":3},"attestation_state":"computed","paper":{"title":"Understanding Dataset Difficulty with $\\mathcal{V}$-Usable Information","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Kawin Ethayarajh, Swabha Swayamdipta, Yejin Choi","submitted_at":"2021-10-16T00:21:42Z","abstract_excerpt":"Estimating the difficulty of a dataset typically involves comparing state-of-the-art models to humans; the bigger the performance gap, the harder the dataset is said to be. However, this comparison provides little understanding of how difficult each instance in a given distribution is, or what attributes make the dataset difficult for a given model. To address these questions, we frame dataset difficulty -- w.r.t. a model $\\mathcal{V}$ -- as the lack of $\\mathcal{V}$-$\\textit{usable information}$ (Xu et al., 2019), where a lower value indicates a more difficult dataset for $\\mathcal{V}$. We fu"},"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":"2110.08420","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-10-16T00:21:42Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"f377b5fddf8f2584771c4f9f6130767732accc680e0c7765b58c97f30114f23b","abstract_canon_sha256":"406804b05e0209ff9b72f8fdf9193e85f9d9bbefcfdde6464dbf068acdb7ac3a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:54:28.689689Z","signature_b64":"H86krPEkHaFUwXrU2AfL7cMxJw2RFhr2yLlFOiaBIqe1OkCggUoxzufGIWNzSF/8JnFN83sHA1hYJt9kEskQDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6816c5cfb7f427858fc29ea7f6d5736cc7423c3a15573ea6c215fcf14e34ed0e","last_reissued_at":"2026-07-05T10:54:28.689158Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:54:28.689158Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Understanding Dataset Difficulty with $\\mathcal{V}$-Usable Information","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Kawin Ethayarajh, Swabha Swayamdipta, Yejin Choi","submitted_at":"2021-10-16T00:21:42Z","abstract_excerpt":"Estimating the difficulty of a dataset typically involves comparing state-of-the-art models to humans; the bigger the performance gap, the harder the dataset is said to be. However, this comparison provides little understanding of how difficult each instance in a given distribution is, or what attributes make the dataset difficult for a given model. To address these questions, we frame dataset difficulty -- w.r.t. a model $\\mathcal{V}$ -- as the lack of $\\mathcal{V}$-$\\textit{usable information}$ (Xu et al., 2019), where a lower value indicates a more difficult dataset for $\\mathcal{V}$. We fu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.08420","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/2110.08420/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":"2110.08420","created_at":"2026-07-05T10:54:28.689221+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.08420v3","created_at":"2026-07-05T10:54:28.689221+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.08420","created_at":"2026-07-05T10:54:28.689221+00:00"},{"alias_kind":"pith_short_12","alias_value":"NALMLT5X6QTY","created_at":"2026-07-05T10:54:28.689221+00:00"},{"alias_kind":"pith_short_16","alias_value":"NALMLT5X6QTYLD6C","created_at":"2026-07-05T10:54:28.689221+00:00"},{"alias_kind":"pith_short_8","alias_value":"NALMLT5X","created_at":"2026-07-05T10:54:28.689221+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.00769","citing_title":"LitBench: A Benchmark and Dataset for Reliable Evaluation of Creative Writing","ref_index":10,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NALMLT5X6QTYLD6CT2T7NVLTNT","json":"https://pith.science/pith/NALMLT5X6QTYLD6CT2T7NVLTNT.json","graph_json":"https://pith.science/api/pith-number/NALMLT5X6QTYLD6CT2T7NVLTNT/graph.json","events_json":"https://pith.science/api/pith-number/NALMLT5X6QTYLD6CT2T7NVLTNT/events.json","paper":"https://pith.science/paper/NALMLT5X"},"agent_actions":{"view_html":"https://pith.science/pith/NALMLT5X6QTYLD6CT2T7NVLTNT","download_json":"https://pith.science/pith/NALMLT5X6QTYLD6CT2T7NVLTNT.json","view_paper":"https://pith.science/paper/NALMLT5X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.08420&json=true","fetch_graph":"https://pith.science/api/pith-number/NALMLT5X6QTYLD6CT2T7NVLTNT/graph.json","fetch_events":"https://pith.science/api/pith-number/NALMLT5X6QTYLD6CT2T7NVLTNT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NALMLT5X6QTYLD6CT2T7NVLTNT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NALMLT5X6QTYLD6CT2T7NVLTNT/action/storage_attestation","attest_author":"https://pith.science/pith/NALMLT5X6QTYLD6CT2T7NVLTNT/action/author_attestation","sign_citation":"https://pith.science/pith/NALMLT5X6QTYLD6CT2T7NVLTNT/action/citation_signature","submit_replication":"https://pith.science/pith/NALMLT5X6QTYLD6CT2T7NVLTNT/action/replication_record"}},"created_at":"2026-07-05T10:54:28.689221+00:00","updated_at":"2026-07-05T10:54:28.689221+00:00"}