diff --git "a/_source/_legacy/dataset_metadata.json" "b/_source/_legacy/dataset_metadata.json" --- "a/_source/_legacy/dataset_metadata.json" +++ "b/_source/_legacy/dataset_metadata.json" @@ -1,4260 +1,3 @@ -{ - "builder_sha256": "bcd3e1d0be6fa9dfbf60381f7407997fc61fc4a7f07e97d3a6728a45f8fdf704", - "candidate_order": "Shuffle all options, including positives; preserve IDs, values and soft labels. Paired decisions with identical option IDs share order.", - "datasets": { - "aegis2": { - "description": "安全性ポリシーに照らしてプロンプトや応答の問題を判定する、コンテンツモデレーションのデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "aegis2", - "statistics": { - "test": { - "cases": 100, - "decisions": 148 - } - }, - "task_types": { - "noul": 148 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/nvidia/Aegis-AI-Content-Safety-Dataset-2.0/blob/d86bb8bedff51d25ac834ab7838f1cc61acb7a2c/README.md", - "name": "nvidia/Aegis-AI-Content-Safety-Dataset-2.0", - "revision": "d86bb8bedff51d25ac834ab7838f1cc61acb7a2c", - "url": "https://huggingface.co/datasets/nvidia/Aegis-AI-Content-Safety-Dataset-2.0" - } - ] - }, - "ag_news": { - "description": "ニュースの文章を、世界・スポーツ・ビジネス・科学技術のカテゴリへ分類するデータセット。Laya向けに変換したinstructionと教師を保持し、元データ由来の分割で型付き判断の学習や評価に利用できます。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "ag_news", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "unknown" - ], - "license_checked_revision": "eb185aade064a813bc0b7f42de02595523103ca4", - "license_lookup_error": null, - "license_source": "https://huggingface.co/datasets/fancyzhx/ag_news/blob/main/README.md", - "name": "fancyzhx/ag_news", - "url": "https://huggingface.co/datasets/fancyzhx/ag_news" - } - ] - }, - "aqua_rat": { - "description": "数量や単位を含む文章問題について、数学的な推論から正しい解答を選ぶデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "duplicate_choice_policy": "exact_choice_text_mass_merge_v1", - "evaluation_suite": null, - "language": [ - "en" - ], - "post_repair_full_statistics": { - "instructions": [ - { - "decisions": 100, - "text": "Solve the mathematical question using the stated quantities and units." - } - ], - "languages": { - "en": 100 - }, - "lengths": { - "docs": null, - "instruction": { - "count": 100, - "max": 70, - "mean": 70.0, - "median": 70, - "min": 70, - "p95": 70 - }, - "options": { - "count": 499, - "max": 26, - "mean": 4.62124248496994, - "median": 3, - "min": 1, - "p95": 13 - }, - "options_per_decision": { - "count": 100, - "max": 5, - "mean": 4.99, - "median": 5, - "min": 4, - "p95": 5 - }, - "query": null, - "state_field:question": { - "count": 100, - "max": 475, - "mean": 188.11, - "median": 176, - "min": 16, - "p95": 342 - }, - "state_json": { - "count": 100, - "max": 490, - "mean": 203.27, - "median": 191, - "min": 31, - "p95": 357 - } - }, - "splits": { - "test": { - "cases": 100, - "decisions": 100, - "labeled_decisions": 100, - "languages": { - "en": 100 - }, - "lengths": { - "docs": null, - "instruction": { - "count": 100, - "max": 70, - "mean": 70.0, - "median": 70, - "min": 70, - "p95": 70 - }, - "options": { - "count": 499, - "max": 26, - "mean": 4.62124248496994, - "median": 3, - "min": 1, - "p95": 13 - }, - "options_per_decision": { - "count": 100, - "max": 5, - "mean": 4.99, - "median": 5, - "min": 4, - "p95": 5 - }, - "query": null, - "state_field:question": { - "count": 100, - "max": 475, - "mean": 188.11, - "median": 176, - "min": 16, - "p95": 342 - }, - "state_json": { - "count": 100, - "max": 490, - "mean": 203.27, - "median": 191, - "min": 31, - "p95": 357 - } - }, - "target_kinds": { - "hard_label": 100 - }, - "task_types": { - "choice": 100 - }, - "unlabeled_decisions": 0 - } - }, - "validation": "all rows: prompt IDs aligned, target IDs aligned, finite nonnegative unit-sum probabilities; generated rows also checked for role/label mapping, option counts, distinct ranking docs and recorded token limits" - }, - "source_subset": "aqua_rat", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/deepmind/aqua_rat/blob/33301c6a050c96af81f63cad5562cb5363e88971/README.md", - "name": "deepmind/aqua_rat", - "revision": "33301c6a050c96af81f63cad5562cb5363e88971", - "url": "https://huggingface.co/datasets/deepmind/aqua_rat" - } - ] - }, - "arc": { - "description": "科学の知識や推論を必要とする多肢選択問題を集めた、質問応答のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "arc", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-4.0" - ], - "license_source": "https://huggingface.co/datasets/allenai/ai2_arc/blob/210d026faf9955653af8916fad021475a3f00453/README.md", - "name": "allenai/ai2_arc", - "revision": "210d026faf9955653af8916fad021475a3f00453", - "url": "https://huggingface.co/datasets/allenai/ai2_arc" - } - ] - }, - "arct": { - "description": "主張と理由の関係を成立させる暗黙の前提を選ぶ、議論理解のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "arct", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/arct/blob/c13f83ea610d0f04c8b6ea50a59339b8204dcd44/README.md", - "name": "tasksource/arct", - "revision": "c13f83ea610d0f04c8b6ea50a59339b8204dcd44", - "url": "https://huggingface.co/datasets/tasksource/arct" - } - ] - }, - "argument_quality": { - "description": "与えられた論題に対する議論の質を評価する、自然言語の論証評価データセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "argument_quality", - "statistics": { - "test": { - "cases": 100, - "decisions": 200 - } - }, - "task_types": { - "choice": 100, - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-3.0" - ], - "license_source": "https://huggingface.co/datasets/ibm-research/argument_quality_ranking_30k/blob/590726b3765b1b90c5e53a17e3b1f77d92d3aa8a/README.md", - "name": "ibm-research/argument_quality_ranking_30k", - "revision": "590726b3765b1b90c5e53a17e3b1f77d92d3aa8a", - "url": "https://huggingface.co/datasets/ibm-research/argument_quality_ranking_30k" - } - ] - }, - "babi_nli": { - "description": "短い物語や事実の記述から、対象の状態や関係について推論するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で��持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "babi_nli", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "bsd" - ], - "license_source": "https://huggingface.co/datasets/tasksource/babi_nli/blob/80b689dc668b50f3640fc5368fd1893d385cb4e1/README.md", - "name": "tasksource/babi_nli", - "revision": "80b689dc668b50f3640fc5368fd1893d385cb4e1", - "url": "https://huggingface.co/datasets/tasksource/babi_nli" - } - ] - }, - "banking77": { - "description": "銀行サービスに関する利用者の問い合わせを、細かな意図カテゴリに分類するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "banking77", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/PolyAI/banking77/blob/90d4e2ee5521c04fc1488f065b8b083658768c57/README.md", - "name": "PolyAI/banking77", - "revision": "90d4e2ee5521c04fc1488f065b8b083658768c57", - "url": "https://huggingface.co/datasets/PolyAI/banking77" - } - ] - }, - "bbq": { - "description": "社会集団に関する文脈付き質問を使い、曖昧さと偏見への依存を調べる評価用データセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "bbq", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/lighteval/bbq_helm/blob/10937569dd76b6f33b2ff3572382ed87458e7856/README.md", - "name": "lighteval/bbq_helm", - "revision": "10937569dd76b6f33b2ff3572382ed87458e7856", - "url": "https://huggingface.co/datasets/lighteval/bbq_helm" - } - ] - }, - "boardgameqa": { - "description": "ルールと事実から結論を判断し、矛盾や例外を伴う推論を扱う質問応答データセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "boardgameqa", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/Boardgame-QA/blob/78e38c3c8df3b4f6de7ae8bd1fc6a8bd1f31be56/README.md", - "name": "tasksource/Boardgame-QA", - "revision": "78e38c3c8df3b4f6de7ae8bd1fc6a8bd1f31be56", - "url": "https://huggingface.co/datasets/tasksource/Boardgame-QA" - } - ] - }, - "canttalk": { - "description": "ユーザーのメッセージが許可された話題の範囲内にあるかを判断する、話題制御のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "canttalk", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/nvidia/CantTalkAboutThis-Topic-Control-Dataset/blob/b015ae8902d8429ce32ebfed9f9935f891d10ed5/README.md", - "name": "nvidia/CantTalkAboutThis-Topic-Control-Dataset", - "revision": "b015ae8902d8429ce32ebfed9f9935f891d10ed5", - "url": "https://huggingface.co/datasets/nvidia/CantTalkAboutThis-Topic-Control-Dataset" - } - ] - }, - "civil_comments": { - "description": "オンラ���ンのコメントに含まれる有害性や関連属性を、人手注釈に基づいて判定するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "civil_comments", - "statistics": { - "test": { - "cases": 100, - "decisions": 700 - } - }, - "task_types": { - "noul": 700 - }, - "upstream_datasets": [ - { - "license": [ - "cc0-1.0" - ], - "license_source": "https://huggingface.co/datasets/google/civil_comments/blob/f2970eb3a55777454c94069077cc8d9b5866312d/README.md", - "name": "google/civil_comments", - "revision": "f2970eb3a55777454c94069077cc8d9b5866312d", - "url": "https://huggingface.co/datasets/google/civil_comments" - } - ] - }, - "cladder": { - "description": "因果関係を記述した問題について、因果推論に基づく回答を評価するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "cladder", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://github.com/causalNLP/cladder/blob/3d2d1169b4b939a09048a6a75956c8972a93cc38/README.md", - "name": "https://github.com/causalNLP/cladder", - "revision": "3d2d1169b4b939a09048a6a75956c8972a93cc38", - "url": "https://github.com/causalNLP/cladder" - } - ] - }, - "clinc": { - "description": "対話システムへの発話を複数のサービス領域にまたがる意図へ分類するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "clinc", - "statistics": { - "test": { - "cases": 151, - "decisions": 151 - } - }, - "task_types": { - "choice": 151 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-3.0" - ], - "license_source": "https://huggingface.co/datasets/clinc/clinc_oos/blob/155b9c710419136e17307b80d0a13e68cd46b4ec/README.md", - "name": "clinc/clinc_oos", - "revision": "155b9c710419136e17307b80d0a13e68cd46b4ec", - "url": "https://huggingface.co/datasets/clinc/clinc_oos" - } - ] - }, - "contract_nli": { - "description": "契約書の内容が指定された命題を支持するかを判断する、法律文書の含意認識データセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "contract_nli", - "statistics": { - "test": { - "cases": 100, - "decisions": 1700 - } - }, - "task_types": { - "choice": 1700 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://github.com/stanfordnlp/contract-nli/blob/eced6528dd3c1d14d73f9a87df8f7bdbc03126f9/README.md", - "name": "https://github.com/stanfordnlp/contract-nli", - "revision": "eced6528dd3c1d14d73f9a87df8f7bdbc03126f9", - "url": "https://github.com/stanfordnlp/contract-nli" - } - ] - }, - "corr2cause": { - "description": "変数間の統計的関係の記述から、因果的な結論が導けるかを判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "corr2cause", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/tasksource/corr2cause/blob/aa0e8b909b8cb7589eb23dad5e9c1cb709e665a3/README.md", - "name": "tasksource/corr2cause", - "revision": "aa0e8b909b8cb7589eb23dad5e9c1cb709e665a3", - "url": "https://huggingface.co/datasets/tasksource/corr2cause" - } - ] - }, - "creak": { - "description": "実世界のエンティティに関する短い主張の真偽を、常識や知識に基づいて判定するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "creak", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/amydeng2000/CREAK/blob/cceb4696560317e920d6512b906263bb425883a1/README.md", - "name": "amydeng2000/CREAK", - "revision": "cceb4696560317e920d6512b906263bb425883a1", - "url": "https://huggingface.co/datasets/amydeng2000/CREAK" - } - ] - }, - "crows_pairs": { - "description": "社会的なステレオタイプに関わる対照文を用いて、モデルの偏りを調べる評価用データセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "crows_pairs", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-4.0" - ], - "license_source": "https://github.com/nyu-mll/crows-pairs/blob/8aaac11c485473159ec9328a65253a5be9a479dc/README.md", - "name": "https://github.com/nyu-mll/crows-pairs", - "revision": "8aaac11c485473159ec9328a65253a5be9a479dc", - "url": "https://github.com/nyu-mll/crows-pairs" - } - ] - }, - "dbpedia": { - "description": "百科事典由来の項目説明を、人物・組織・場所などのカテゴリへ分類するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "dbpedia", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-3.0" - ], - "license_source": "https://huggingface.co/datasets/fancyzhx/dbpedia_14/blob/9abd46cf7fc8b4c64290f26993c540b92aa145ac/README.md", - "name": "fancyzhx/dbpedia_14", - "revision": "9abd46cf7fc8b4c64290f26993c540b92aa145ac", - "url": "https://huggingface.co/datasets/fancyzhx/dbpedia_14" - } - ] - }, - "defeasible_nli": { - "description": "追加情報によって推論の支持が強まるか弱まるかを扱う、撤回可能な推論のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "defeasible_nli", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0", - "cc-by-sa-4.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/defeasible-nli/blob/7c4a57df9d8de5c36d4e9caa977907b5e8469c4f/README.md", - "name": "tasksource/defeasible-nli", - "revision": "7c4a57df9d8de5c36d4e9caa977907b5e8469c4f", - "url": "https://huggingface.co/datasets/tasksource/defeasible-nli" - } - ] - }, - "enron_spam": { - "description": "電子メールの内容を読み、通常メールと迷惑メールを識別するデータセット。Laya向けに変換したinstructionと教師を保持し、元データ由来の分割で型付き判断の学習や評価に利用できます。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "enron_spam", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": "unknown", - "license_checked_revision": "1916f66c89d52221ae33eb57d44498b4f3a5df22", - "license_lookup_error": null, - "license_source": "https://huggingface.co/datasets/SetFit/enron_spam/blob/main/README.md", - "name": "SetFit/enron_spam", - "url": "https://huggingface.co/datasets/SetFit/enron_spam" - } - ] - }, - "esci": { - "description": "商品検索クエリと商品情報の関係を、適合・代替などの区分で判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "esci", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/esci/blob/8113b17a5d4099e20243282c926f1bc1a08a4d13/README.md", - "name": "tasksource/esci", - "revision": "8113b17a5d4099e20243282c926f1bc1a08a4d13", - "url": "https://huggingface.co/datasets/tasksource/esci" - } - ] - }, - "ethics": { - "description": "日常的な行為や判断を倫理の観点から評価する、複数の課題を含むデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "ethics", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 18, - "noul": 82 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/hendrycks/ethics/blob/b8b47c589f8bee77175b8648e5497278b68da48a/README.md", - "name": "hendrycks/ethics", - "revision": "b8b47c589f8bee77175b8648e5497278b68da48a", - "url": "https://huggingface.co/datasets/hendrycks/ethics" - } - ] - }, - "ethos": { - "description": "オンラインの発言に含まれるヘイト表現やその属性を判定するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "ethos", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/SetFit/ethos_binary/blob/3e20849175072aa95845981f27938ebeb4eb93e0/README.md", - "name": "SetFit/ethos_binary", - "revision": "3e20849175072aa95845981f27938ebeb4eb93e0", - "url": "https://huggingface.co/datasets/SetFit/ethos_binary" - } - ] - }, - "few_nerd": { - "description": "文中のエンティティを細かな意味カテゴリで識別する、固有表現認識由来のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "few_nerd", - "statistics": { - "test": { - "cases": 100, - "decisions": 358 - } - }, - "task_types": { - "choice": 358 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-4.0" - ], - "license_source": "https://huggingface.co/datasets/DFKI-SLT/few-nerd/blob/205f3e9c9f3577ea2561d43f2f62dc249ab92d5b/README.md", - "name": "DFKI-SLT/few-nerd", - "revision": "205f3e9c9f3577ea2561d43f2f62dc249ab92d5b", - "url": "https://huggingface.co/datasets/DFKI-SLT/few-nerd" - } - ] - }, - "fol_nli": { - "description": "一階述語論理に関係する文の意味や帰結を扱う、論理的含意判断のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "fol_nli", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/FOL-nli/blob/6b8a2ec01b226ed871fbde4677a31146ac5f370c/README.md", - "name": "tasksource/FOL-nli", - "revision": "6b8a2ec01b226ed871fbde4677a31146ac5f370c", - "url": "https://huggingface.co/datasets/tasksource/FOL-nli" - } - ] - }, - "followir_core17": { - "description": "検索queryと文書の関連性を、個別のinstructionに従ってYes/Noで判断するFollowIRデータです。学習例と3評価suiteを統合し、評価では同一文書への指示変更前後の2判断を保持します。", - "evaluation_policy": "Keep original/changed judgments in one case. Report both-correct accuracy separately for changed-label and unchanged-label pairs, by evaluation_suite. This derived Noul evaluation is not official p-MRR.", - "evaluation_suite": "core17", - "language": [ - "en" - ], - "license_scope": "HF card licenses are recorded per source; original evaluation corpus redistribution terms are separate.", - "source_subset": "followir", - "statistics": { - "test": { - "cases": 100, - "decisions": 200 - } - }, - "task_types": { - "noul": 200 - }, - "teacher_provenance": "Train: GPT-3.5 synthetic documents/labels filtered by Mistral; preserve source label, retain Mistral score only as metadata. Test: explicitly judged qrels > 0.", - "upstream_datasets": [ - { - "license": "mit", - "name": "mteb/Core17InstructionRetrieval", - "revision": "67d8733dd0691f08f34ed9dda5941578cba8d91c", - "url": "https://huggingface.co/datasets/mteb/Core17InstructionRetrieval" - } - ] - }, - "followir_news21": { - "description": "検索queryと文書の関連性を、個別のinstructionに従ってYes/Noで判断するFollowIRデータです。学習例と3評価suiteを統合し、評価では同一文書への指示変更前後の2判断を保持します。", - "evaluation_policy": "Keep original/changed judgments in one case. Report both-correct accuracy separately for changed-label and unchanged-label pairs, by evaluation_suite. This derived Noul evaluation is not official p-MRR.", - "evaluation_suite": "news21", - "language": [ - "en" - ], - "license_scope": "HF card licenses are recorded per source; original evaluation corpus redistribution terms are separate.", - "source_subset": "followir", - "statistics": { - "test": { - "cases": 100, - "decisions": 200 - } - }, - "task_types": { - "noul": 200 - }, - "teacher_provenance": "Train: GPT-3.5 synthetic documents/labels filtered by Mistral; preserve source label, retain Mistral score only as metadata. Test: explicitly judged qrels > 0.", - "upstream_datasets": [ - { - "license": "mit", - "name": "mteb/News21InstructionRetrieval", - "revision": "c6fffeb9cdd95c1ffa81c62c687b48b56352872b", - "url": "https://huggingface.co/datasets/mteb/News21InstructionRetrieval" - } - ] - }, - "followir_robust04": { - "description": "検索queryと文書の関連性を、個別のinstructionに従ってYes/Noで判断するFollowIRデータです。学習例と3評価suiteを統合し、評価では同一文書への指示変更前後の2判断を保持します。", - "evaluation_policy": "Keep original/changed judgments in one case. Report both-correct accuracy separately for changed-label and unchanged-label pairs, by evaluation_suite. This derived Noul evaluation is not official p-MRR.", - "evaluation_suite": "robust04", - "language": [ - "en" - ], - "license_scope": "HF card licenses are recorded per source; original evaluation corpus redistribution terms are separate.", - "source_subset": "followir", - "statistics": { - "test": { - "cases": 100, - "decisions": 200 - } - }, - "task_types": { - "noul": 200 - }, - "teacher_provenance": "Train: GPT-3.5 synthetic documents/labels filtered by Mistral; preserve source label, retain Mistral score only as metadata. Test: explicitly judged qrels > 0.", - "upstream_datasets": [ - { - "license": "mit", - "name": "mteb/Robust04InstructionRetrieval", - "revision": "0495a5cb69aa8fa5bca81bd56ac248911c19eeb6", - "url": "https://huggingface.co/datasets/mteb/Robust04InstructionRetrieval" - } - ] - }, - "go_emotions": { - "description": "短いオンライン投稿から、細かな感情カテゴリの該当を判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "go_emotions", - "statistics": { - "test": { - "cases": 100, - "decisions": 2800 - } - }, - "task_types": { - "noul": 2800 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/google-research-datasets/go_emotions/blob/add492243ff905527e67aeb8b80c082af02207c3/README.md", - "name": "google-research-datasets/go_emotions", - "revision": "add492243ff905527e67aeb8b80c082af02207c3", - "url": "https://huggingface.co/datasets/google-research-datasets/go_emotions" - } - ] - }, - "gretel_pii": { - "description": "文章中の個人識別情報を扱い、対象となる情報種別を判定するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "gretel_pii", - "statistics": { - "test": { - "cases": 100, - "decisions": 353 - } - }, - "task_types": { - "choice": 353 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/gretelai/gretel-pii-masking-en-v1/blob/e06eb1499ca8d54470f085021cd8e54f9efac7fd/README.md", - "name": "gretelai/gretel-pii-masking-en-v1", - "revision": "e06eb1499ca8d54470f085021cd8e54f9efac7fd", - "url": "https://huggingface.co/datasets/gretelai/gretel-pii-masking-en-v1" - } - ] - }, - "gsm8k": { - "description": "小学校水準の算数文章題を対象に、複数段階の計算を必要とする解答を判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "gsm8k", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/openai/gsm8k/blob/740312add88f781978c0658806c59bc2815b9866/README.md", - "name": "openai/gsm8k", - "revision": "740312add88f781978c0658806c59bc2815b9866", - "url": "https://huggingface.co/datasets/openai/gsm8k" - } - ] - }, - "hans": { - "description": "語の重なりなどの表面的な手掛かりに頼らず、文間の含意を判断できるか調べるデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "hans", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/tasksource/hans/blob/8f107047d34fdfb57abfde38ec5196d63f7dee33/README.md", - "name": "tasksource/hans", - "revision": "8f107047d34fdfb57abfde38ec5196d63f7dee33", - "url": "https://huggingface.co/datasets/tasksource/hans" - } - ] - }, - "hatecheck": { - "description": "ヘイト表現の機能的なテストケースを使い、検出器の判断を評価するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "hatecheck", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/Paul/hatecheck/blob/9d2ac89df04254e5c427bcc8d61b6d6c83a1f59b/README.md", - "name": "Paul/hatecheck", - "revision": "9d2ac89df04254e5c427bcc8d61b6d6c83a1f59b", - "url": "https://huggingface.co/datasets/Paul/hatecheck" - } - ] - }, - "hh_rlhf": { - "description": "対話への応答を有用性や無害性の観点で比較する、人手選好のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "hh_rlhf", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/Anthropic/hh-rlhf/blob/09be8c5bbc57cb3887f3a9732ad6aa7ec602a1fa/README.md", - "name": "Anthropic/hh-rlhf", - "revision": "09be8c5bbc57cb3887f3a9732ad6aa7ec602a1fa", - "url": "https://huggingface.co/datasets/Anthropic/hh-rlhf" - } - ] - }, - "hwu64": { - "description": "複数の領域にまたがるアシスタント向け発話から、その意図を分類するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "hwu64", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-3.0" - ], - "license_source": "https://huggingface.co/datasets/FastFit/hwu_64/blob/9fb5bff1e37c4c5a6d46c1ff6f286b6d3f362222/README.md", - "name": "FastFit/hwu_64", - "revision": "9fb5bff1e37c4c5a6d46c1ff6f286b6d3f362222", - "url": "https://huggingface.co/datasets/FastFit/hwu_64" - } - ] - }, - "impli": { - "description": "慣用的・比喩的な表現を含む文の関係から、意味的な含意を調べるデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "impli", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://github.com/UKPLab/acl2022-impli/blob/d43c272c895c773b311018d51ab65f592c6699bb/README.md", - "name": "https://github.com/UKPLab/acl2022-impli", - "revision": "d43c272c895c773b311018d51ab65f592c6699bb", - "url": "https://github.com/UKPLab/acl2022-impli" - } - ] - }, - "lexcomp": { - "description": "複合語の意味を候補の言い換えが正しく説明するか、文脈を踏まえて判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "lexcomp", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/lexcomp-nc-relation/blob/4f7fa5d487d71ac1543bc7849993b5dfc2e361ac/README.md", - "name": "tasksource/lexcomp-nc-relation", - "revision": "4f7fa5d487d71ac1543bc7849993b5dfc2e361ac", - "url": "https://huggingface.co/datasets/tasksource/lexcomp-nc-relation" - } - ] - }, - "logical_entailment": { - "description": "提示された論理的な前提から結論が導けるかを判断する、含意推論のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "logical_entailment", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/logical-entailment/blob/5275a96cea765dc3bbfd41e6034485791704ed69/README.md", - "name": "tasksource/logical-entailment", - "revision": "5275a96cea765dc3bbfd41e6034485791704ed69", - "url": "https://huggingface.co/datasets/tasksource/logical-entailment" - } - ] - }, - "lonli": { - "description": "さまざまな論理現象を制御した文対で、含意・矛盾などの判断を評価するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "lonli", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/tasksource/lonli/blob/f17e55b8f60da2f32720c1a16e4b6149145c662f/README.md", - "name": "tasksource/lonli", - "revision": "f17e55b8f60da2f32720c1a16e4b6149145c662f", - "url": "https://huggingface.co/datasets/tasksource/lonli" - } - ] - }, - "massive": { - "description": "多領域のアシスタント発話から意図を判定するデータセットの、採択済み英語部分。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "massive", - "statistics": { - "test": { - "cases": 100, - "decisions": 200 - } - }, - "task_types": { - "choice": 200 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/AmazonScience/massive/blob/ff6bd8e4b27c3543e4f8fe2108f32bb95a6f8740/README.md", - "name": "AmazonScience/massive", - "revision": "ff6bd8e4b27c3543e4f8fe2108f32bb95a6f8740", - "url": "https://huggingface.co/datasets/AmazonScience/massive" - } - ] - }, - "miqa": { - "description": "比喩的な意味の理解を必要とする質問応答を通して、推論を調べるデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "miqa", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0", - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/miqa/blob/b0b4a19860eea1740f04a7836612ec239e08e838/README.md", - "name": "tasksource/miqa", - "revision": "b0b4a19860eea1740f04a7836612ec239e08e838", - "url": "https://huggingface.co/datasets/tasksource/miqa" - } - ] - }, - "mtop": { - "description": "対話アシスタントへの要求を、意図などの意味構造に沿って判定するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "mtop", - "statistics": { - "test": { - "cases": 100, - "decisions": 200 - } - }, - "task_types": { - "choice": 200 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-4.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/mtop/blob/9fa95557ba2809ad05943880eee5231b4f2463fb/README.md", - "name": "tasksource/mtop", - "revision": "9fa95557ba2809ad05943880eee5231b4f2463fb", - "url": "https://huggingface.co/datasets/tasksource/mtop" - } - ] - }, - "nlsat": { - "description": "自然言語で記述された論理的条件について、整合性や成立可能性を判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "nlsat", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/nlsat/blob/1c7e1a3afbb62462ee979d8d77521ebbfcef0255/README.md", - "name": "tasksource/nlsat", - "revision": "1c7e1a3afbb62462ee979d8d77521ebbfcef0255", - "url": "https://huggingface.co/datasets/tasksource/nlsat" - } - ] - }, - "openbookqa": { - "description": "基礎的な科学の事実と常識を組み合わせ、多肢選択の質問に答えるデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "openbookqa", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/allenai/openbookqa/blob/388097ea7776314e93a529163e0fea805b8a6454/README.md", - "name": "allenai/openbookqa", - "revision": "388097ea7776314e93a529163e0fea805b8a6454", - "url": "https://huggingface.co/datasets/allenai/openbookqa" - } - ] - }, - "patent_similarity": { - "description": "特許に関係する語句の組について、文脈を踏まえた意味的な類似性を評価するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "patent_similarity", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "score": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/patent-phrase-similarity/blob/ff1868796e44ac9ec02e870b447fde6c5bf2d6e4/README.md", - "name": "tasksource/patent-phrase-similarity", - "revision": "ff1868796e44ac9ec02e870b447fde6c5bf2d6e4", - "url": "https://huggingface.co/datasets/tasksource/patent-phrase-similarity" - } - ] - }, - "paws": { - "description": "語の重なりが大きい文対でも、実際に同じ意味を表すかを見分けるデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "paws", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "other" - ], - "license_source": "https://huggingface.co/datasets/google-research-datasets/paws/blob/161ece9501cf0a11f3e48bd356eaa82de46d6a09/README.md", - "name": "google-research-datasets/paws", - "revision": "161ece9501cf0a11f3e48bd356eaa82de46d6a09", - "url": "https://huggingface.co/datasets/google-research-datasets/paws" - } - ] - }, - "plane": { - "description": "カテゴリに関する記述から、指定されたカテゴリ推論が妥当かを判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "plane", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/PLANE-ood/blob/5ca649ccef611c6d2b453d4b34d25477a63bd9e4/README.md", - "name": "tasksource/PLANE-ood", - "revision": "5ca649ccef611c6d2b453d4b34d25477a63bd9e4", - "url": "https://huggingface.co/datasets/tasksource/PLANE-ood" - } - ] - }, - "poem_sentiment": { - "description": "詩のテキストに表れる感情の極性を分類する、文章感情分析のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "poem_sentiment", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/google-research-datasets/poem_sentiment/blob/685b95a2787a869b7bae6c4480810f57fe23b48e/README.md", - "name": "google-research-datasets/poem_sentiment", - "revision": "685b95a2787a869b7bae6c4480810f57fe23b48e", - "url": "https://huggingface.co/datasets/google-research-datasets/poem_sentiment" - } - ] - }, - "qasper": { - "description": "学術論文の内容に関する質問と根拠を扱う、科学文書の読解データセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "qasper", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/allenai/qasper/blob/fdc9d8214fbab5dd782958601db4d678e6934a54/README.md", - "name": "allenai/qasper", - "revision": "fdc9d8214fbab5dd782958601db4d678e6934a54", - "url": "https://huggingface.co/datasets/allenai/qasper" - } - ] - }, - "quartz": { - "description": "量の増減などの定性的な関係を文章から読み取り、質問に答えるデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "quartz", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/allenai/quartz/blob/28c1dbb56caf81799296cb17892fa73402e23464/README.md", - "name": "allenai/quartz", - "revision": "28c1dbb56caf81799296cb17892fa73402e23464", - "url": "https://huggingface.co/datasets/allenai/quartz" - } - ] - }, - "robust_lr": { - "description": "論理的な前提や表現の変更に対する推論の頑健性を調べる評価用データセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "robust_lr", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/tasksource/robustLR/blob/369ce16fbe4467f46916efe3730d04fb607c0391/README.md", - "name": "tasksource/robustLR", - "revision": "369ce16fbe4467f46916efe3730d04fb607c0391", - "url": "https://huggingface.co/datasets/tasksource/robustLR" - } - ] - }, - "ruletaker": { - "description": "自然言語で書かれたルールと事実から、指定された結論が導けるかを判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "ruletaker", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/ruletaker/blob/a3e0880baeb6ec3d478f4c4d85afe04b21b6cf7f/README.md", - "name": "tasksource/ruletaker", - "revision": "a3e0880baeb6ec3d478f4c4d85afe04b21b6cf7f", - "url": "https://huggingface.co/datasets/tasksource/ruletaker" - } - ] - }, - "scicite": { - "description": "学術論文における引用の文脈から、引用の目的を分類するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "scicite", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/scicite/blob/6b569c1b7045f187cccc94c38af91faa755178ae/README.md", - "name": "tasksource/scicite", - "revision": "6b569c1b7045f187cccc94c38af91faa755178ae", - "url": "https://huggingface.co/datasets/tasksource/scicite" - } - ] - }, - "scitail": { - "description": "科学に関する前提文と仮説文の組について、含意が成立するかを判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "scitail", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/allenai/scitail/blob/0cc4353235b289165dfde1c7c5d1be983f99ce44/README.md", - "name": "allenai/scitail", - "revision": "0cc4353235b289165dfde1c7c5d1be983f99ce44", - "url": "https://huggingface.co/datasets/allenai/scitail" - } - ] - }, - "scone": { - "description": "自然言語の推論に影響する複数の言語現象を扱い、文対の意味関係を判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "scone", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc0-1.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/scone/blob/505ed270f9e9865c6cfab1e222d804fedd837127/README.md", - "name": "tasksource/scone", - "revision": "505ed270f9e9865c6cfab1e222d804fedd837127", - "url": "https://huggingface.co/datasets/tasksource/scone" - } - ] - }, - "sdoh_nli": { - "description": "健康の社会的決定要因に関係する記述から、指定された内容が支持されるかを判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "sdoh_nli", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-4.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/SDOH-NLI/blob/75d4cc5f67ecb3f3c44cf5a3fadc3c9579f2f148/README.md", - "name": "tasksource/SDOH-NLI", - "revision": "75d4cc5f67ecb3f3c44cf5a3fadc3c9579f2f148", - "url": "https://huggingface.co/datasets/tasksource/SDOH-NLI" - } - ] - }, - "sgd": { - "description": "複数サービスの対話とスキーマに基づき、利用者の意図や要求を判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "sgd", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-4.0" - ], - "license_source": "https://github.com/google-research-datasets/dstc8-schema-guided-dialogue/blob/e852981ae34990f4358979625854259302feaa78/README.md", - "name": "https://github.com/google-research-datasets/dstc8-schema-guided-dialogue", - "revision": "e852981ae34990f4358979625854259302feaa78", - "url": "https://github.com/google-research-datasets/dstc8-schema-guided-dialogue" - } - ] - }, - "snli": { - "description": "前提文と仮説文の組を、含意・矛盾・中立の関係で判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "snli", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-4.0" - ], - "license_source": "https://huggingface.co/datasets/stanfordnlp/snli/blob/cdb5c3d5eed6ead6e5a341c8e56e669bb666725b/README.md", - "name": "stanfordnlp/snli", - "revision": "cdb5c3d5eed6ead6e5a341c8e56e669bb666725b", - "url": "https://huggingface.co/datasets/stanfordnlp/snli" - } - ] - }, - "spartqa": { - "description": "空間的な配置を記述した文章を読み、物体間の関係について質問に答えるデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "spartqa", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/tasksource/spartqa-mchoice/blob/d29ba456044649c3c2fe29c5ff429c3e6d80a97b/README.md", - "name": "tasksource/spartqa-mchoice", - "revision": "d29ba456044649c3c2fe29c5ff429c3e6d80a97b", - "url": "https://huggingface.co/datasets/tasksource/spartqa-mchoice" - } - ] - }, - "stepgame": { - "description": "複数の位置関係をつなぎ合わせ、対象間の空間的な関係を推論するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "stepgame", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/tasksource/stepgame/blob/3ba235d4a7056cf8b062a57aa489f7bc14bff229/README.md", - "name": "tasksource/stepgame", - "revision": "3ba235d4a7056cf8b062a57aa489f7bc14bff229", - "url": "https://huggingface.co/datasets/tasksource/stepgame" - } - ] - }, - "synthetic_relevance__nanobeir__NanoArguAna": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 25, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoArguAna", - "source_tasks": { - "NanoArguAna": { - "bibtex": "@inproceedings{wachsmuth2018arguana,\n title = {Retrieval of the Best Counterargument without Prior Topic Knowledge},\n author = {Wachsmuth, Henning and Syed, Shahbaz and Stein, Benno},\n booktitle = {Proceedings of ACL},\n year = {2018},\n url = {https://aclanthology.org/P18-1023/},\n doi = {10.18653/v1/P18-1023}\n}\n", - "category": "natural_language", - "citation_keys": [ - "wachsmuth2018arguana" - ], - "description": "A compact ArguAna split where each query is an argument and relevant documents are strong counterarguments. It tests whether retrieval models capture argumentative stance and semantic opposition beyond lexical overlap. This split is the compact NanoArguAna subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 3635, - "max_chars": 6673, - "mean_chars": 1011.7914718019257, - "median_chars": 904.0, - "min_chars": 70 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 100.0 - }, - "sample_count": 3635 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 2164, - "mean_chars": 1201.78, - "median_chars": 1170.5, - "min_chars": 504 - }, - "references": [ - { - "authors": [ - "Henning Wachsmuth", - "Shahbaz Syed", - "Benno Stein" - ], - "doi": "10.18653/v1/P18-1023", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "Retrieval of the Best Counterargument without Prior Topic Knowledge", - "url": "https://aclanthology.org/P18-1023/", - "year": 2018 - } - ], - "short_description": "Argument retrieval over counterargument pairs." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoArguAna", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoClimateFEVER": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 68, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoClimateFEVER", - "source_tasks": { - "NanoClimateFEVER": { - "bibtex": "@misc{diggelmann2021climatefever,\n title = {{CLIMATE-FEVER}: A Dataset for Verification of Real-World Climate Claims},\n author = {Diggelmann, Thomas and Boyd-Graber, Jordan and Bulian, Jannis and Ciaramita, Massimiliano and Leippold, Markus},\n year = {2021},\n url = {https://arxiv.org/abs/2012.00614},\n doi = {10.48550/arXiv.2012.00614}\n}\n", - "category": "natural_language", - "citation_keys": [ - "diggelmann2021climatefever" - ], - "description": "A compact Climate-FEVER split for retrieving evidence documents for real-world climate claims. It measures claim-evidence matching in a fact-checking setting with climate-domain terminology. This split is the compact NanoClimateFEVER subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 3408, - "max_chars": 6619, - "mean_chars": 1619.531690140845, - "median_chars": 1461.0, - "min_chars": 33 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 98.21 - }, - "sample_count": 3408 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 265, - "mean_chars": 128.4, - "median_chars": 124.0, - "min_chars": 38 - }, - "references": [ - { - "authors": [ - "Thomas Diggelmann", - "Jordan Boyd-Graber", - "Jannis Bulian", - "Massimiliano Ciaramita", - "Markus Leippold" - ], - "doi": "10.48550/arXiv.2012.00614", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CLIMATE-FEVER: A Dataset for Verification of Real-World Climate Claims", - "url": "https://arxiv.org/abs/2012.00614", - "year": 2021 - } - ], - "short_description": "Climate claim evidence retrieval." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoClimateFEVER", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoDBPedia": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 67, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoDBPedia", - "source_tasks": { - "NanoDBPedia": { - "bibtex": "@inproceedings{hasibi2017dbpedia,\n title = {{DBpedia}-Entity V2: A Test Collection for Entity Search},\n author = {Hasibi, Faegheh and Nikolaev, Fedor and Xiong, Chenyan and Balog, Krisztian and Bratsberg, Svein Erik and Kotov, Alexander and Callan, Jamie},\n booktitle = {Proceedings of SIGIR},\n year = {2017},\n url = {https://doi.org/10.1145/3077136.3080751},\n doi = {10.1145/3077136.3080751}\n}\n", - "category": "natural_language", - "citation_keys": [ - "hasibi2017dbpedia" - ], - "description": "A compact DBpedia entity-search split where natural-language queries must retrieve relevant DBpedia entity documents. It emphasizes entity disambiguation and matching descriptions to knowledge-base pages. This split is the compact NanoDBPedia subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 6045, - "max_chars": 1390, - "mean_chars": 336.30669975186106, - "median_chars": 369.0, - "min_chars": 1 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 98.577 - }, - "sample_count": 6045 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 98.0, - "nl": 2.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 63, - "mean_chars": 33.1, - "median_chars": 33.5, - "min_chars": 8 - }, - "references": [ - { - "authors": [ - "Faegheh Hasibi", - "Fedor Nikolaev", - "Chenyan Xiong", - "Krisztian Balog", - "Svein Erik Bratsberg", - "Alexander Kotov", - "Jamie Callan" - ], - "doi": "10.1145/3077136.3080751", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "DBpedia-Entity V2: A Test Collection for Entity Search", - "url": "https://doi.org/10.1145/3077136.3080751", - "year": 2017 - } - ], - "short_description": "Entity retrieval over DBpedia descriptions." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoDBPedia", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoFEVER": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 14, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoFEVER", - "source_tasks": { - "NanoFEVER": { - "bibtex": "@inproceedings{thorne2018fever,\n title = {{FEVER}: a Large-scale Dataset for Fact Extraction and Verification},\n author = {Thorne, James and Vlachos, Andreas and Christodoulopoulos, Christos and Mittal, Arpit},\n booktitle = {Proceedings of NAACL-HLT},\n year = {2018},\n url = {https://aclanthology.org/N18-1074/},\n doi = {10.18653/v1/N18-1074}\n}\n", - "category": "natural_language", - "citation_keys": [ - "thorne2018fever" - ], - "description": "A compact FEVER split for retrieving Wikipedia evidence relevant to factual claims. It evaluates fact-checking retrieval where claims must be matched to evidence-bearing documents. This split is the compact NanoFEVER subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 4996, - "max_chars": 8491, - "mean_chars": 1228.7119695756605, - "median_chars": 1041.0, - "min_chars": 25 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 97.438, - "ru": 0.701 - }, - "sample_count": 4996 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 83, - "mean_chars": 45.42, - "median_chars": 43.0, - "min_chars": 17 - }, - "references": [ - { - "authors": [ - "James Thorne", - "Andreas Vlachos", - "Christos Christodoulopoulos", - "Arpit Mittal" - ], - "doi": "10.18653/v1/N18-1074", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "FEVER: a Large-scale Dataset for Fact Extraction and Verification", - "url": "https://aclanthology.org/N18-1074/", - "year": 2018 - } - ], - "short_description": "Wikipedia evidence retrieval for claims." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoFEVER", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoFiQA2018": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 21, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoFiQA2018", - "source_tasks": { - "NanoFiQA2018": { - "bibtex": "@inproceedings{maia2018fiqa,\n title = {{WWW}'18 Open Challenge: Financial Opinion Mining and Question Answering},\n author = {Maia, Macedo and Handschuh, Siegfried and Freitas, Andre and Davis, Brian and McDermott, Ross and Zarrouk, Manel and Balahur, Alexandra},\n booktitle = {The 2018 Web Conference Companion},\n year = {2018},\n url = {https://doi.org/10.1145/3184558.3192301},\n doi = {10.1145/3184558.3192301}\n}\n", - "category": "natural_language", - "citation_keys": [ - "maia2018fiqa" - ], - "description": "A compact FiQA-2018 retrieval split focused on finance-domain questions and opinion-bearing answers. It tests retrieval under specialized financial terminology and user question phrasing. This split is the compact NanoFiQA2018 subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 4571, - "max_chars": 10506, - "mean_chars": 904.9466199956246, - "median_chars": 655.0, - "min_chars": 30 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 100.0 - }, - "sample_count": 4571 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 97, - "mean_chars": 58.52, - "median_chars": 55.0, - "min_chars": 18 - }, - "references": [ - { - "authors": [ - "Macedo Maia", - "Siegfried Handschuh", - "Andre Freitas", - "Brian Davis", - "Ross McDermott", - "Manel Zarrouk", - "Alexandra Balahur" - ], - "doi": "10.1145/3184558.3192301", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "WWW'18 Open Challenge: Financial Opinion Mining and Question Answering", - "url": "https://doi.org/10.1145/3184558.3192301", - "year": 2018 - } - ], - "short_description": "Financial opinion question answering retrieval." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoFiQA2018", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoHotpotQA": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 28, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoHotpotQA", - "source_tasks": { - "NanoHotpotQA": { - "bibtex": "@inproceedings{yang2018hotpotqa,\n title = {{HotpotQA}: A Dataset for Diverse, Explainable Multi-hop Question Answering},\n author = {Yang, Zhilin and Qi, Peng and Zhang, Saizheng and Bengio, Yoshua and Cohen, William W. and Salakhutdinov, Ruslan and Manning, Christopher D.},\n booktitle = {Proceedings of EMNLP},\n year = {2018},\n url = {https://aclanthology.org/D18-1259/},\n doi = {10.18653/v1/D18-1259}\n}\n", - "category": "natural_language", - "citation_keys": [ - "yang2018hotpotqa" - ], - "description": "A compact HotpotQA retrieval split where questions require finding relevant Wikipedia evidence. It emphasizes multi-hop question intent and paragraph-level evidence retrieval. This split is the compact NanoHotpotQA subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 5090, - "max_chars": 2079, - "mean_chars": 349.6349705304519, - "median_chars": 299.0, - "min_chars": 24 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 98.9 - }, - "sample_count": 5090 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 184, - "mean_chars": 88.34, - "median_chars": 82.5, - "min_chars": 37 - }, - "references": [ - { - "authors": [ - "Zhilin Yang", - "Peng Qi", - "Saizheng Zhang", - "Yoshua Bengio", - "William W. Cohen", - "Ruslan Salakhutdinov", - "Christopher D. Manning" - ], - "doi": "10.18653/v1/D18-1259", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering", - "url": "https://aclanthology.org/D18-1259/", - "year": 2018 - } - ], - "short_description": "Multi-hop Wikipedia question retrieval." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoHotpotQA", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoMSMARCO": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 12, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoMSMARCO", - "source_tasks": { - "NanoMSMARCO": { - "bibtex": "@article{nguyen2016msmarco,\n title = {{MS MARCO}: A Human Generated Machine Reading Comprehension Dataset},\n author = {Nguyen, Tri and Rosenberg, Mir and Song, Xia and Gao, Jianfeng and Tiwary, Saurabh and Majumder, Rangan and Deng, Li},\n journal = {arXiv preprint arXiv:1611.09268},\n year = {2016},\n url = {https://arxiv.org/abs/1611.09268},\n doi = {10.48550/arXiv.1611.09268}\n}\n", - "category": "natural_language", - "citation_keys": [ - "nguyen2016msmarco" - ], - "description": "A compact MS MARCO passage-retrieval split using real web-search style questions. It evaluates matching short user queries to answer-bearing passages. This split is the compact NanoMSMARCO subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 5043, - "max_chars": 990, - "mean_chars": 330.159825500694, - "median_chars": 298.0, - "min_chars": 32 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 99.861 - }, - "sample_count": 5043 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "de": 2.0, - "en": 98.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 101, - "mean_chars": 32.22, - "median_chars": 29.0, - "min_chars": 13 - }, - "references": [ - { - "authors": [ - "Tri Nguyen", - "Mir Rosenberg", - "Xia Song", - "Jianfeng Gao", - "Saurabh Tiwary", - "Rangan Majumder", - "Li Deng" - ], - "doi": "10.48550/arXiv.1611.09268", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "MS MARCO: A Human Generated MAchine Reading COmprehension Dataset", - "url": "https://arxiv.org/abs/1611.09268", - "year": 2016 - } - ], - "short_description": "Web passage retrieval from MS MARCO queries." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoMSMARCO", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoNFCorpus": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 72, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoNFCorpus", - "source_tasks": { - "NanoNFCorpus": { - "bibtex": "@inproceedings{boteva2016nfcorpus,\n title = {A Full-Text Learning to Rank Dataset for Medical Information Retrieval},\n author = {Boteva, Vera and Gholipour Ghalandari, Demian and Sokolov, Artem and Riezler, Stefan},\n booktitle = {Advances in Information Retrieval},\n year = {2016},\n url = {https://www.cl.uni-heidelberg.de/~riezler/publications/papers/ECIR2016.pdf},\n doi = {10.1007/978-3-319-30671-1_58}\n}\n", - "category": "natural_language", - "citation_keys": [ - "boteva2016nfcorpus" - ], - "description": "A compact NFCorpus split for medical information retrieval. Queries are natural-language health information needs, and relevant documents come from medically oriented full-text content. This split is the compact NanoNFCorpus subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 2953, - "max_chars": 9939, - "mean_chars": 1512.7301049779885, - "median_chars": 1532.0, - "min_chars": 90 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 99.932 - }, - "sample_count": 2953 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "de": 2.0, - "en": 94.0, - "fr": 2.0, - "it": 2.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 53, - "mean_chars": 21.04, - "median_chars": 16.5, - "min_chars": 4 - }, - "references": [ - { - "authors": [ - "Vera Boteva", - "Demian Gholipour Ghalandari", - "Artem Sokolov", - "Stefan Riezler" - ], - "doi": "10.1007/978-3-319-30671-1_58", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "A Full-Text Learning to Rank Dataset for Medical Information Retrieval", - "url": "https://www.cl.uni-heidelberg.de/~riezler/publications/papers/ECIR2016.pdf", - "year": 2016 - } - ], - "short_description": "Medical information retrieval from NFCorpus." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoNFCorpus", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoNQ": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 9, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoNQ", - "source_tasks": { - "NanoNQ": { - "bibtex": "@article{kwiatkowski2019nq,\n title = {Natural Questions: A Benchmark for Question Answering Research},\n author = {Kwiatkowski, Tom and Palomaki, Jennimaria and Redfield, Olivia and Collins, Michael and Parikh, Ankur and Alberti, Chris and Epstein, Danielle and Polosukhin, Illia and Devlin, Jacob and Lee, Kenton and others},\n journal = {Transactions of the Association for Computational Linguistics},\n volume = {7},\n pages = {452--466},\n year = {2019},\n url = {https://aclanthology.org/Q19-1026/},\n doi = {10.1162/tacl_a_00276}\n}\n", - "category": "natural_language", - "citation_keys": [ - "kwiatkowski2019nq" - ], - "description": "A compact Natural Questions retrieval split built from real user questions. It tests whether models retrieve Wikipedia documents or passages that contain answers to information-seeking questions. This split is the compact NanoNQ subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 5035, - "max_chars": 6138, - "mean_chars": 525.5958291956306, - "median_chars": 443.0, - "min_chars": 1 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 99.464 - }, - "sample_count": 5035 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 83, - "mean_chars": 47.04, - "median_chars": 42.5, - "min_chars": 32 - }, - "references": [ - { - "authors": [ - "Tom Kwiatkowski", - "Jennimaria Palomaki", - "Olivia Redfield", - "Michael Collins", - "Ankur Parikh", - "Chris Alberti", - "Danielle Epstein", - "Illia Polosukhin", - "Jacob Devlin", - "Kenton Lee", - "Kristina Toutanova", - "Llion Jones", - "Matthew Kelcey", - "Ming-Wei Chang", - "Andrew M. Dai", - "Jakob Uszkoreit", - "Quoc Le", - "Slav Petrov" - ], - "doi": "10.1162/tacl_a_00276", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "Natural Questions: A Benchmark for Question Answering Research", - "url": "https://aclanthology.org/Q19-1026/", - "year": 2019 - } - ], - "short_description": "Natural Questions Wikipedia retrieval." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoNQ", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoQuoraRetrieval": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 26, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoQuoraRetrieval", - "source_tasks": { - "NanoQuoraRetrieval": { - "bibtex": "@misc{quora2017questionpairs,\n title = {Quora Question Pairs},\n author = {DataCanary and hilfialkaff and Jiang, Lili and Risdal, Meg and Dandekar, Nikhil and tomtung},\n year = {2017},\n url = {https://kaggle.com/competitions/quora-question-pairs}\n}\n", - "category": "natural_language", - "citation_keys": [ - "quora2017questionpairs" - ], - "description": "A compact Quora duplicate-question retrieval split. Queries are questions and relevant documents are semantically duplicate questions, testing paraphrase-level matching. This split is the compact NanoQuoraRetrieval subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 5046, - "max_chars": 332, - "mean_chars": 54.808164883075705, - "median_chars": 47.0, - "min_chars": 2 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 99.941 - }, - "sample_count": 5046 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 139, - "mean_chars": 47.96, - "median_chars": 43.5, - "min_chars": 19 - }, - "references": [ - { - "authors": [ - "DataCanary", - "hilfialkaff", - "Lili Jiang", - "Meg Risdal", - "Nikhil Dandekar", - "tomtung" - ], - "is_paper": false, - "source_confidence": "probably_correct", - "title": "Quora Question Pairs", - "url": "https://kaggle.com/competitions/quora-question-pairs", - "year": 2017 - } - ], - "short_description": "Duplicate question retrieval from Quora pairs." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoQuoraRetrieval", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoSCIDOCS": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 75, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoSCIDOCS", - "source_tasks": { - "NanoSCIDOCS": { - "bibtex": "@inproceedings{cohan2020specter,\n title = {{SPECTER}: Document-level Representation Learning using Citation-informed Transformers},\n author = {Cohan, Arman and Feldman, Sergey and Beltagy, Iz and Downey, Doug and Weld, Daniel S.},\n booktitle = {Proceedings of ACL},\n year = {2020},\n url = {https://aclanthology.org/2020.acl-main.207/},\n doi = {10.18653/v1/2020.acl-main.207}\n}\n", - "category": "natural_language", - "citation_keys": [ - "cohan2020specter" - ], - "description": "A compact SCIDOCS retrieval split over scientific papers. It evaluates retrieval of related academic documents using citation-informed scholarly relevance signals. This split is the compact NanoSCIDOCS subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 1866, - "max_chars": 10000, - "mean_chars": 1093.8322615219722, - "median_chars": 996.0, - "min_chars": 28 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 99.893 - }, - "sample_count": 1866 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 143, - "mean_chars": 72.78, - "median_chars": 71.5, - "min_chars": 38 - }, - "references": [ - { - "authors": [ - "Arman Cohan", - "Sergey Feldman", - "Iz Beltagy", - "Doug Downey", - "Daniel S. Weld" - ], - "doi": "10.18653/v1/2020.acl-main.207", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "SPECTER: Document-level Representation Learning using Citation-informed Transformers", - "url": "https://aclanthology.org/2020.acl-main.207/", - "year": 2020 - } - ], - "short_description": "Scientific document retrieval from SCIDOCS." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoSCIDOCS", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoSciFact": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 19, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoSciFact", - "source_tasks": { - "NanoSciFact": { - "bibtex": "@inproceedings{wadden2020scifact,\n title = {Fact or Fiction: Verifying Scientific Claims},\n author = {Wadden, David and Lin, Shanchuan and Lo, Kyle and Wang, Lucy Lu and van Zuylen, Madeleine and Cohan, Arman and Hajishirzi, Hannaneh},\n booktitle = {Proceedings of EMNLP},\n year = {2020},\n url = {https://aclanthology.org/2020.emnlp-main.609/},\n doi = {10.18653/v1/2020.emnlp-main.609}\n}\n", - "category": "natural_language", - "citation_keys": [ - "wadden2020scifact" - ], - "description": "A compact SciFact split for retrieving scientific abstracts that support or refute claims. It tests evidence retrieval in research literature with scientific terminology. This split is the compact NanoSciFact subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 2919, - "max_chars": 10000, - "mean_chars": 1431.2343268242548, - "median_chars": 1344.0, - "min_chars": 260 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 100.0 - }, - "sample_count": 2919 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 200, - "mean_chars": 95.8, - "median_chars": 92.5, - "min_chars": 37 - }, - "references": [ - { - "authors": [ - "David Wadden", - "Shanchuan Lin", - "Kyle Lo", - "Lucy Lu Wang", - "Madeleine van Zuylen", - "Arman Cohan", - "Hannaneh Hajishirzi" - ], - "doi": "10.18653/v1/2020.emnlp-main.609", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "Fact or Fiction: Verifying Scientific Claims", - "url": "https://aclanthology.org/2020.emnlp-main.609/", - "year": 2020 - } - ], - "short_description": "Scientific claim evidence retrieval." - } - }, - "statistics": { - "test": { - "cases": 400, - "decisions": 400 - } - }, - "task_types": { - "score": 400 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoSciFact", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanobeir__NanoTouche2020": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 46, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanobeir__NanoTouche2020", - "source_tasks": { - "NanoTouche2020": { - "bibtex": "@inproceedings{bondarenko2020touche,\n title = {Overview of {Touché} 2020: Argument Retrieval},\n author = {Bondarenko, Alexander and Fröbe, Maik and Beloucif, Meriem and Gienapp, Lukas and Ajjour, Yamen and Panchenko, Alexander and Biemann, Chris and Stein, Benno and Wachsmuth, Henning and Potthast, Martin and Hagen, Matthias},\n booktitle = {Experimental IR Meets Multilinguality, Multimodality, and Interaction},\n year = {2020},\n url = {https://doi.org/10.1007/978-3-030-58219-7_26},\n doi = {10.1007/978-3-030-58219-7_26}\n}\n@dataset{potthast2022touche20,\n title = {{Touche20}-Argument-Retrieval-for-Controversial-Questions},\n author = {Potthast, Martin and Gienapp, Lukas and Wachsmuth, Henning and Hagen, Matthias\n and Fröbe, Maik and Bondarenko, Alexander and Ajjour, Yamen and Stein, Benno},\n year = {2022},\n url = {https://doi.org/10.5281/zenodo.6862281},\n doi = {10.5281/zenodo.6862281}\n}\n", - "category": "natural_language", - "citation_keys": [ - "bondarenko2020touche", - "potthast2022touche20" - ], - "description": "A compact Touche 2020 argument-retrieval split. Queries are controversial questions, and relevant documents provide arguments that address the topic. This split is the compact NanoTouche2020 subset used by NanoBEIR-en.", - "document_text_stats": { - "count": 5745, - "max_chars": 37100, - "mean_chars": 2142.56953872933, - "median_chars": 989.0, - "min_chars": 3 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 99.426 - }, - "sample_count": 5745 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 49 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 49, - "max_chars": 83, - "mean_chars": 43.42857142857143, - "median_chars": 40.0, - "min_chars": 16 - }, - "references": [ - { - "authors": [ - "Alexander Bondarenko", - "Maik Fröbe", - "Meriem Beloucif", - "Lukas Gienapp", - "Yamen Ajjour", - "Alexander Panchenko", - "Chris Biemann", - "Benno Stein", - "Henning Wachsmuth", - "Martin Potthast", - "Matthias Hagen" - ], - "doi": "10.1007/978-3-030-58219-7_26", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "Overview of Touch�� 2020: Argument Retrieval", - "url": "https://doi.org/10.1007/978-3-030-58219-7_26", - "year": 2020 - }, - { - "authors": [ - "Martin Potthast", - "Lukas Gienapp", - "Henning Wachsmuth", - "Matthias Hagen", - "Maik Fröbe", - "Alexander Bondarenko", - "Yamen Ajjour", - "Benno Stein" - ], - "doi": "10.5281/zenodo.6862281", - "is_paper": false, - "source_confidence": "probably_correct", - "title": "Touche20-Argument-Retrieval-for-Controversial-Questions", - "url": "https://doi.org/10.5281/zenodo.6862281", - "year": 2022 - } - ], - "short_description": "Argument retrieval for controversial questions." - } - }, - "statistics": { - "test": { - "cases": 392, - "decisions": 392 - } - }, - "task_types": { - "score": 392 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoTouche2020", - "revision": "d3962aa8efe48ed79044c5e155b848982667b4ba", - "url": "https://huggingface.co/datasets/hakari-bench/NanoBEIR-en" - } - ] - }, - "synthetic_relevance__nanocoir": { - "description": "Nano検索候補からquery単位で8文書を抽出した評価専用の5段階関連性データ。DeepSeekの採点に元positiveの最低3点補正を適用し、元タスク・出典・元スコアを保持します。", - "evaluation_policy": "test only; one document per case, eight per source query; keep all sampled query groups", - "evaluation_suite": null, - "language": [ - "en" - ], - "positive_correction": { - "corrected_cases": 22, - "policy": "source positive raw 0/1/2 -> 3" - }, - "source_subset": "synthetic_relevance__nanocoir", - "source_tasks": { - "NanoApps": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n\n@inproceedings{hendrycks2021apps,\n title = {Measuring Coding Challenge Competence With APPS},\n author = {Hendrycks, Dan and Basart, Steven and Kadavath, Saurav and Mazeika, Mantas and Arora, Akul and Guo, Ethan and Burns, Collin and Puranik, Samir and He, Horace and Song, Dawn and Steinhardt, Jacob},\n booktitle = {Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks},\n year = {2021},\n url = {https://datasets-benchmarks-proceedings.neurips.cc/paper_files/paper/2021/hash/c24cd76e1ce41366a4bbe8a49b02a028-Abstract-round2.html}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir", - "hendrycks2021apps" - ], - "description": "Uses programming problem statements as queries and candidate solution code as documents, testing retrieval for competitive-programming style tasks.", - "document_text_stats": { - "count": 8754, - "max_chars": 289048, - "mean_chars": 573.1170893305917, - "median_chars": 331.0, - "min_chars": 5 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "ceb": 3.438, - "en": 88.2, - "fa": 2.81, - "ja": 0.64, - "ru": 1.062, - "zh": 1.382 - }, - "sample_count": 8754 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 98.5, - "ru": 1.5 - }, - "sample_count": 200 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 200, - "max_chars": 4088, - "mean_chars": 1675.415, - "median_chars": 1589.0, - "min_chars": 411 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - }, - { - "authors": [ - "Dan Hendrycks", - "Steven Basart", - "Saurav Kadavath", - "Mantas Mazeika", - "Akul Arora", - "Ethan Guo", - "Collin Burns", - "Samir Puranik", - "Horace He", - "Dawn Song", - "Jacob Steinhardt" - ], - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "Measuring Coding Challenge Competence With APPS", - "url": "https://datasets-benchmarks-proceedings.neurips.cc/paper_files/paper/2021/hash/c24cd76e1ce41366a4bbe8a49b02a028-Abstract-round2.html", - "year": 2021 - } - ], - "short_description": "APPS coding challenge retrieval." - }, - "NanoCodeFeedbackMT": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n\n@misc{zheng2024opencodeinterpreter,\n title = {OpenCodeInterpreter: Integrating Code Generation with Execution and Refinement},\n author = {Zheng, Tianyu and Zhang, Ge and Shen, Tianhao and Liu, Xueling and Lin, Bill Yuchen and Fu, Jie and Chen, Wenhu and Yue, Xiang},\n year = {2024},\n eprint = {2402.14658},\n archivePrefix = {arXiv},\n url = {https://arxiv.org/abs/2402.14658},\n doi = {10.48550/arXiv.2402.14658}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir", - "zheng2024opencodeinterpreter" - ], - "description": "Retrieves the next relevant code-assistant response from long multi-turn programming conversations adapted from OpenCodeInterpreter-style feedback data.", - "document_text_stats": { - "count": 10000, - "max_chars": 9402, - "mean_chars": 1468.1565, - "median_chars": 1373.0, - "min_chars": 5 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 99.34 - }, - "sample_count": 10000 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 200 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 200, - "max_chars": 21256, - "mean_chars": 4468.625, - "median_chars": 3662.5, - "min_chars": 285 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - }, - { - "authors": [ - "Tianyu Zheng", - "Ge Zhang", - "Tianhao Shen", - "Xueling Liu", - "Bill Yuchen Lin", - "Jie Fu", - "Wenhu Chen", - "Xiang Yue" - ], - "doi": "10.48550/arXiv.2402.14658", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "OpenCodeInterpreter: Integrating Code Generation with Execution and Refinement", - "url": "https://arxiv.org/abs/2402.14658", - "year": 2024 - } - ], - "short_description": "Multi-turn code feedback retrieval." - }, - "NanoCodeFeedbackST": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir" - ], - "description": "Retrieves code-assistant responses for single-turn programming instructions from the CoIR CodeFeedback-ST task.", - "document_text_stats": { - "count": 10000, - "max_chars": 11353, - "mean_chars": 1538.7276, - "median_chars": 1407.0, - "min_chars": 1 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 98.08 - }, - "sample_count": 10000 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 200 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 200, - "max_chars": 3406, - "mean_chars": 730.515, - "median_chars": 598.5, - "min_chars": 57 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - } - ], - "short_description": "Single-turn code feedback retrieval." - }, - "NanoCodeSearchNet": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n\n@article{husain2019codesearchnet,\n title = {{CodeSearchNet} Challenge: Evaluating the State of Semantic Code Search},\n author = {Husain, Hamel and Wu, Ho-Hsiang and Gazit, Tiferet and Allamanis, Miltiadis and Brockschmidt, Marc},\n journal = {arXiv preprint arXiv:1909.09436},\n year = {2019},\n url = {https://arxiv.org/abs/1909.09436},\n doi = {10.48550/arXiv.1909.09436}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir", - "husain2019codesearchnet" - ], - "description": "Uses CodeSearchNet pairs to retrieve the natural-language summary or documentation associated with a code snippet.", - "document_text_stats": { - "count": 10000, - "max_chars": 2610, - "mean_chars": 86.072, - "median_chars": 65.0, - "min_chars": 8 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 94.34, - "es": 1.5, - "ru": 1.0, - "uk": 0.58, - "zh": 0.51 - }, - "sample_count": 10000 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "ceb": 0.5, - "en": 85.5, - "es": 1.0, - "hu": 0.5, - "nl": 0.5, - "pt": 5.5, - "ru": 2.5, - "tr": 1.5, - "uk": 0.5, - "zh": 2.0 - }, - "sample_count": 200 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 200, - "max_chars": 3345, - "mean_chars": 636.265, - "median_chars": 434.0, - "min_chars": 83 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - }, - { - "authors": [ - "Hamel Husain", - "Ho-Hsiang Wu", - "Tiferet Gazit", - "Miltiadis Allamanis", - "Marc Brockschmidt" - ], - "doi": "10.48550/arXiv.1909.09436", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CodeSearchNet Challenge: Evaluating the State of Semantic Code Search", - "url": "https://arxiv.org/abs/1909.09436", - "year": 2019 - } - ], - "short_description": "CodeSearchNet code-summary retrieval." - }, - "NanoCodeSearchNetCCR": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir" - ], - "description": "Retrieves continuation code for partial function contexts in CoIR's CodeSearchNet-CCR code-to-code setting.", - "document_text_stats": { - "count": 10000, - "max_chars": 1849, - "mean_chars": 158.4226, - "median_chars": 99.0, - "min_chars": 23 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "ar": 0.74, - "de": 1.74, - "en": 78.37, - "es": 3.42, - "fa": 1.09, - "fr": 0.57, - "it": 1.26, - "ja": 2.0, - "pt": 0.68, - "ru": 2.48, - "uk": 1.35, - "zh": 1.71 - }, - "sample_count": 10000 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "ceb": 0.5, - "en": 86.0, - "es": 1.0, - "hu": 0.5, - "pt": 5.5, - "ru": 2.5, - "tr": 1.5, - "uk": 0.5, - "zh": 2.0 - }, - "sample_count": 200 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 200, - "max_chars": 2240, - "mean_chars": 372.82, - "median_chars": 237.0, - "min_chars": 47 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - } - ], - "short_description": "CodeSearchNet code-context retrieval." - }, - "NanoCodeTransOceanContest": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n\n@inproceedings{yan2023codetransocean,\n title = {{CodeTransOcean}: A Comprehensive Multilingual Benchmark for Code Translation},\n author = {Yan, Weixiang and Tian, Yuchen and Li, Yunzhe and Chen, Qian and Wang, Wen},\n booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2023},\n pages = {5067--5089},\n year = {2023},\n address = {Singapore},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2023.findings-emnlp.337/},\n doi = {10.18653/v1/2023.findings-emnlp.337}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir", - "yan2023codetransocean" - ], - "description": "Uses CodeTransOcean contest data to retrieve semantically equivalent competitive-programming solutions across programming languages.", - "document_text_stats": { - "count": 1008, - "max_chars": 10876, - "mean_chars": 1528.720238095238, - "median_chars": 1118.5, - "min_chars": 16 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 88.095, - "es": 8.433, - "ja": 1.587, - "uk": 1.091 - }, - "sample_count": 1008 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "de": 0.5, - "en": 98.0, - "es": 0.5, - "pt": 1.0 - }, - "sample_count": 200 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 200, - "max_chars": 4142, - "mean_chars": 1009.56, - "median_chars": 755.5, - "min_chars": 7 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - }, - { - "authors": [ - "Weixiang Yan", - "Yuchen Tian", - "Yunzhe Li", - "Qian Chen", - "Wen Wang" - ], - "doi": "10.18653/v1/2023.findings-emnlp.337", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CodeTransOcean: A Comprehensive Multilingual Benchmark for Code Translation", - "url": "https://aclanthology.org/2023.findings-emnlp.337/", - "year": 2023 - } - ], - "short_description": "Cross-language contest code retrieval." - }, - "NanoCodeTransOceanDL": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n\n@inproceedings{yan2023codetransocean,\n title = {{CodeTransOcean}: A Comprehensive Multilingual Benchmark for Code Translation},\n author = {Yan, Weixiang and Tian, Yuchen and Li, Yunzhe and Chen, Qian and Wang, Wen},\n booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2023},\n pages = {5067--5089},\n year = {2023},\n address = {Singapore},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2023.findings-emnlp.337/},\n doi = {10.18653/v1/2023.findings-emnlp.337}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir", - "yan2023codetransocean" - ], - "description": "Uses CodeTransOcean-DL examples to retrieve equivalent Python deep-learning code across machine-learning frameworks.", - "document_text_stats": { - "count": 266, - "max_chars": 8468, - "mean_chars": 1644.9924812030076, - "median_chars": 1314.0, - "min_chars": 57 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 97.744, - "es": 1.88 - }, - "sample_count": 266 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 94.0, - "es": 2.0, - "pl": 2.0, - "ru": 2.0 - }, - "sample_count": 50 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 50, - "max_chars": 8285, - "mean_chars": 2153.8, - "median_chars": 1440.0, - "min_chars": 376 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - }, - { - "authors": [ - "Weixiang Yan", - "Yuchen Tian", - "Yunzhe Li", - "Qian Chen", - "Wen Wang" - ], - "doi": "10.18653/v1/2023.findings-emnlp.337", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CodeTransOcean: A Comprehensive Multilingual Benchmark for Code Translation", - "url": "https://aclanthology.org/2023.findings-emnlp.337/", - "year": 2023 - } - ], - "short_description": "Deep-learning framework code retrieval." - }, - "NanoCosQA": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n\n@inproceedings{huang2021cosqa,\n title = {{CoSQA}: 20,000+ Web Queries for Code Search and Question Answering},\n author = {Huang, Junjie and Tang, Duyu and Shou, Linjun and Gong, Ming and Xu, Ke and Jiang, Daxin and Zhou, Ming and Duan, Nan},\n booktitle = {Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)},\n pages = {5690--5700},\n year = {2021},\n address = {Online},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2021.acl-long.442/},\n doi = {10.18653/v1/2021.acl-long.442}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir", - "huang2021cosqa" - ], - "description": "Uses CoSQA web queries to retrieve relevant Python code, testing short natural-language code search with lexical and semantic gaps.", - "document_text_stats": { - "count": 6267, - "max_chars": 6395, - "mean_chars": 307.6144885910324, - "median_chars": 258.0, - "min_chars": 87 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 98.341 - }, - "sample_count": 6267 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "cs": 1.5, - "en": 95.0, - "es": 0.5, - "fr": 0.5, - "it": 0.5, - "kn": 0.5, - "nl": 0.5, - "ru": 0.5, - "sv": 0.5 - }, - "sample_count": 200 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 200, - "max_chars": 73, - "mean_chars": 36.1, - "median_chars": 35.0, - "min_chars": 18 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - }, - { - "authors": [ - "Junjie Huang", - "Duyu Tang", - "Linjun Shou", - "Ming Gong", - "Ke Xu", - "Daxin Jiang", - "Ming Zhou", - "Nan Duan" - ], - "doi": "10.18653/v1/2021.acl-long.442", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoSQA: 20,000+ Web Queries for Code Search and Question Answering", - "url": "https://aclanthology.org/2021.acl-long.442/", - "year": 2021 - } - ], - "short_description": "Web-query-to-code retrieval." - }, - "NanoStackOverflowQA": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir" - ], - "description": "Retrieves relevant StackOverflow-style answers that may mix prose and code for real programming questions.", - "document_text_stats": { - "count": 10000, - "max_chars": 46027, - "mean_chars": 1218.0589, - "median_chars": 730.0, - "min_chars": 29 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 99.3 - }, - "sample_count": 10000 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 200 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 200, - "max_chars": 15121, - "mean_chars": 1361.805, - "median_chars": 908.5, - "min_chars": 61 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - } - ], - "short_description": "StackOverflow code QA retrieval." - }, - "NanoSyntheticText2SQL": { - "bibtex": "@inproceedings{li2025coir,\n title = {{CoIR}: A Comprehensive Benchmark for Code Information Retrieval Models},\n author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},\n booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},\n pages = {22074--22091},\n year = {2025},\n address = {Vienna, Austria},\n publisher = {Association for Computational Linguistics},\n url = {https://aclanthology.org/2025.acl-long.1072/},\n doi = {10.18653/v1/2025.acl-long.1072}\n}\n\n@software{gretel2024syntheticsql,\n title = {{Synthetic-Text-To-SQL}: A synthetic dataset for training language models to generate SQL queries from natural language prompts},\n author = {Meyer, Yev and Emadi, Marjan and Nathawani, Dhruv and Ramaswamy, Lipika and Boyd, Kendrick and Van Segbroeck, Maarten and Grossman, Matthew and Mlocek, Piotr and Newberry, Drew},\n month = {April},\n year = {2024},\n url = {https://huggingface.co/datasets/gretelai/synthetic-text-to-sql}\n}\n", - "category": "code", - "citation_keys": [ - "li2025coir", - "gretel2024syntheticsql" - ], - "description": "Retrieves SQL statements for natural-language database questions using the Gretel synthetic text-to-SQL dataset adapted in CoIR.", - "document_text_stats": { - "count": 10000, - "max_chars": 730, - "mean_chars": 130.6048, - "median_chars": 111.0, - "min_chars": 20 - }, - "language": "en", - "language_detection": { - "detector": "fast-langdetect", - "document": { - "languages": { - "en": 99.06 - }, - "sample_count": 10000 - }, - "main_language_percent": 10.0, - "min_language_percent": 0.5, - "query": { - "languages": { - "en": 100.0 - }, - "sample_count": 200 - } - }, - "languages": [ - "en" - ], - "query_text_stats": { - "count": 200, - "max_chars": 188, - "mean_chars": 102.935, - "median_chars": 99.0, - "min_chars": 49 - }, - "references": [ - { - "authors": [ - "Xiangyang Li", - "Kuicai Dong", - "Yi Quan Lee", - "Wei Xia", - "Hao Zhang", - "Xinyi Dai", - "Yasheng Wang", - "Ruiming Tang" - ], - "doi": "10.18653/v1/2025.acl-long.1072", - "is_paper": true, - "source_confidence": "definitive_paper_link", - "title": "CoIR: A Comprehensive Benchmark for Code Information Retrieval Models", - "url": "https://aclanthology.org/2025.acl-long.1072/", - "year": 2025 - }, - { - "authors": [ - "Yev Meyer", - "Marjan Emadi", - "Dhruv Nathawani", - "Lipika Ramaswamy", - "Kendrick Boyd", - "Maarten Van Segbroeck", - "Matthew Grossman", - "Piotr Mlocek", - "Drew Newberry" - ], - "is_paper": false, - "source_confidence": "probably_correct", - "title": "Synthetic-Text-To-SQL: A synthetic dataset for training language models to generate SQL queries from natural language prompts", - "url": "https://huggingface.co/datasets/gretelai/synthetic_text_to_sql", - "year": 2024 - } - ], - "short_description": "Text-to-SQL retrieval." - } - }, - "statistics": { - "test": { - "cases": 800, - "decisions": 800 - } - }, - "task_types": { - "score": 800 - }, - "upstream_datasets": [ - { - "license": "unknown", - "name": "NanoApps", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - }, - { - "license": "unknown", - "name": "NanoCodeFeedbackMT", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - }, - { - "license": "unknown", - "name": "NanoCodeFeedbackST", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - }, - { - "license": "unknown", - "name": "NanoCodeSearchNet", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - }, - { - "license": "unknown", - "name": "NanoCodeSearchNetCCR", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - }, - { - "license": "unknown", - "name": "NanoCodeTransOceanContest", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - }, - { - "license": "unknown", - "name": "NanoCodeTransOceanDL", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - }, - { - "license": "unknown", - "name": "NanoCosQA", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - }, - { - "license": "unknown", - "name": "NanoStackOverflowQA", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - }, - { - "license": "unknown", - "name": "NanoSyntheticText2SQL", - "revision": "f6fa3c5c630d51870e260d00ebccdb32abb0afb6", - "url": "https://huggingface.co/datasets/hakari-bench/NanoCoIR" - } - ] - }, - "temporal_nli": { - "description": "時間に関する表現や出来事の関係を読み、文間の含意を判断するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "temporal_nli", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/temporal-nli/blob/d5cedbbdb9f1e7591569ebaf7cf1dd238f0b624b/README.md", - "name": "tasksource/temporal-nli", - "revision": "d5cedbbdb9f1e7591569ebaf7cf1dd238f0b624b", - "url": "https://huggingface.co/datasets/tasksource/temporal-nli" - } - ] - }, - "tracie": { - "description": "出来事の時間的な関係を文脈から推論する、時間常識のデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "tracie", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/tasksource/tracie/blob/4967d8aaa06a4ea97a88bf77071d66be0140bc45/README.md", - "name": "tasksource/tracie", - "revision": "4967d8aaa06a4ea97a88bf77071d66be0140bc45", - "url": "https://huggingface.co/datasets/tasksource/tracie" - } - ] - }, - "typed_decisions": { - "description": "共有された入力状態について、複数の型付き判断を行うデータセット。Laya向けに変換したinstructionと教師を保持し、元データ由来の分割で型付き判断の学習や評価に利用できます。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "typed_decisions", - "statistics": { - "test": { - "cases": 100, - "decisions": 500 - } - }, - "task_types": { - "choice": 148, - "noul": 152, - "score": 200 - }, - "upstream_datasets": [ - { - "license": "apache-2.0", - "license_checked_revision": "f7a2487edd7a043a5441a5e9ccc7fe5ddbd9ebe8", - "license_lookup_error": null, - "license_source": "https://huggingface.co/datasets/LocalLLaMA/typed-decisions/blob/main/README.md", - "name": "LocalLLaMA/typed-decisions", - "url": "https://huggingface.co/datasets/LocalLLaMA/typed-decisions" - } - ] - }, - "ud_ewt": { - "description": "英語のウェブ文章に付けられた言語学的注釈を、型付き判断に変換したデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "ud_ewt", - "statistics": { - "test": { - "cases": 100, - "decisions": 1509 - } - }, - "task_types": { - "choice": 1509 - }, - "upstream_datasets": [ - { - "license": [ - "cc-by-sa-4.0" - ], - "license_source": "https://huggingface.co/datasets/universal-dependencies/universal_dependencies/blob/c3fed78df6b9f8cd27ed24cc7142e9eccf1e6703/README.md", - "name": "universal-dependencies/universal_dependencies", - "revision": "c3fed78df6b9f8cd27ed24cc7142e9eccf1e6703", - "url": "https://huggingface.co/datasets/universal-dependencies/universal_dependencies" - } - ] - }, - "winobias": { - "description": "職業や代名詞の文脈で、性別に関する偏見への依存を調べる評価用データセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "winobias", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://github.com/uclanlp/corefBias/blob/0bce984dd081bbc10b0622f326727a024c607895/README.md", - "name": "https://github.com/uclanlp/corefBias", - "revision": "0bce984dd081bbc10b0622f326727a024c607895", - "url": "https://github.com/uclanlp/corefBias" - } - ] - }, - "winowhy": { - "description": "照応解決の理由の妥当性を判断し、常識推論の説明を調べる評価用データセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "winowhy", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "noul": 100 - }, - "upstream_datasets": [ - { - "license": [ - "mit" - ], - "license_source": "https://huggingface.co/datasets/tasksource/winowhy/blob/b7732098d32396fb9f690e9d50cb7f1758861305/README.md", - "name": "tasksource/winowhy", - "revision": "b7732098d32396fb9f690e9d50cb7f1758861305", - "url": "https://huggingface.co/datasets/tasksource/winowhy" - } - ] - }, - "wiqa": { - "description": "手順や過程への変更が、その結果にどう影響するかを推論するデータセット。元の注釈をChoice・Noul・Scoreの該当形式で保持し、instructionと学習・評価用の分割を収録しています。", - "evaluation_suite": null, - "language": [ - "en" - ], - "source_subset": "wiqa", - "statistics": { - "test": { - "cases": 100, - "decisions": 100 - } - }, - "task_types": { - "choice": 100 - }, - "upstream_datasets": [ - { - "license": [ - "apache-2.0" - ], - "license_source": "https://huggingface.co/datasets/allenai/wiqa/blob/8dda4b5e237452fb939b39326f68ad6607e75ab6/README.md", - "name": "allenai/wiqa", - "revision": "8dda4b5e237452fb939b39326f68ad6607e75ab6", - "url": "https://huggingface.co/datasets/allenai/wiqa" - } - ] - } - }, - "description": "S1MB (System One Mosaic Benchmark) is a mosaic-style benchmark that aggregates dozens of specialized benchmarks to approximate System One capabilities, rather than directly measuring generalization.", - "description_ja": "S1MBは、多数の専門ベンチマークを組み合わせ、System Oneの能力を近似的に評価するベンチマークです。未知のタスクや分布への汎化能力を直接測定するものではありません。", - "duplicate_choice_repair": { - "implementation_sha256": "6d24d686e2a0da4b12c48f17cd4880e0adf47c9d970974d3eea6822d40781a7c", - "policy": "exact_choice_text_mass_merge_v1", - "source": "output/S1B-dataset", - "subsets": { - "aqua_rat": { - "new_cross_split_input_overlaps": {}, - "splits": { - "test": { - "cases": 100, - "changed_case_ids": [ - "aqua_rat:raw:test:193" - ], - "changed_cases": 1, - "conflicting_target_cases_repaired": 1, - "excluded_all_identical_cases": 0, - "excluded_case_ids": [], - "original_cases": 100, - "removed_options": 1 - } - } - } - }, - "timestamp": "20260926-155845" - }, - "exceptions_file": "sampling-exceptions.json", - "followir_policy": "Separate configs per evaluation_suite", - "full_name": "System One Mosaic Benchmark", - "name": "S1MB", - "rename_history": [ - { - "data_changed": false, - "date": "2026-09-26", - "path": "output/S1MB-dataset", - "previous_name": "S1B-dataset", - "previous_path": "output/S1B-dataset" - }, - { - "data_changed": false, - "date": "2026-09-26", - "path": "output/S1MB-dataset", - "previous_name": "MS1B", - "previous_path": "output/MS1B-dataset" - } - ], - "samples_per_role": 100, - "sampling": "balanced-v1: single-axis Noul/fixed Choice equal quotas; multi-axis rare-label case coverage; fixed classes may exceed cap; soft labels preserved. FollowIR unchanged; Nano groups retained. Test only.", - "sampling_helper_sha256": "d35369e689c9a48d3d064c5b1ed8e281550ef062b05b0e89d200dff1fa6ab1de", - "sampling_policy_version": "balanced-v1", - "seed": 42, - "source_manifest_sha256": "b04e8f32400e07c3e5e58efd4100694cff1ad8382329d1c3215a70508ae7d41b", - "source_metadata_sha256": "89c06c646d6d25dc1bd5d3f122952ad9e90d4c524fcc3426283559c12ade456b", - "source_release": "/home/hotchpotch/src/github.com/hotchpotch/system-ichi-exp/output/bekko-system-one-dataset-v1", - "totals": { - "cases": 12843, - "configs": 82, - "decisions": 20711 - } -} +version https://git-lfs.github.com/spec/v1 +oid sha256:8226983930bd6049bd5d602821ce74ee72aabf6034be6a445a69b5bbc04a3a4e +size 165544