Datasets:

ought
/

raft

@@ -1,677 +0,0 @@
-{
-    "ade_corpus_v2": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Sentence": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "ADE-related",
-                    "not ADE-related"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "ade_corpus_v2",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 7602,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 709087,
-                "num_examples": 5000,
-                "dataset_name": null
-            }
-        },
-        "download_size": 445823,
-        "dataset_size": 716689,
-        "size_in_bytes": 1162512
-    },
-    "banking_77": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Query": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "Refund_not_showing_up",
-                    "activate_my_card",
-                    "age_limit",
-                    "apple_pay_or_google_pay",
-                    "atm_support",
-                    "automatic_top_up",
-                    "balance_not_updated_after_bank_transfer",
-                    "balance_not_updated_after_cheque_or_cash_deposit",
-                    "beneficiary_not_allowed",
-                    "cancel_transfer",
-                    "card_about_to_expire",
-                    "card_acceptance",
-                    "card_arrival",
-                    "card_delivery_estimate",
-                    "card_linking",
-                    "card_not_working",
-                    "card_payment_fee_charged",
-                    "card_payment_not_recognised",
-                    "card_payment_wrong_exchange_rate",
-                    "card_swallowed",
-                    "cash_withdrawal_charge",
-                    "cash_withdrawal_not_recognised",
-                    "change_pin",
-                    "compromised_card",
-                    "contactless_not_working",
-                    "country_support",
-                    "declined_card_payment",
-                    "declined_cash_withdrawal",
-                    "declined_transfer",
-                    "direct_debit_payment_not_recognised",
-                    "disposable_card_limits",
-                    "edit_personal_details",
-                    "exchange_charge",
-                    "exchange_rate",
-                    "exchange_via_app",
-                    "extra_charge_on_statement",
-                    "failed_transfer",
-                    "fiat_currency_support",
-                    "get_disposable_virtual_card",
-                    "get_physical_card",
-                    "getting_spare_card",
-                    "getting_virtual_card",
-                    "lost_or_stolen_card",
-                    "lost_or_stolen_phone",
-                    "order_physical_card",
-                    "passcode_forgotten",
-                    "pending_card_payment",
-                    "pending_cash_withdrawal",
-                    "pending_top_up",
-                    "pending_transfer",
-                    "pin_blocked",
-                    "receiving_money",
-                    "request_refund",
-                    "reverted_card_payment?",
-                    "supported_cards_and_currencies",
-                    "terminate_account",
-                    "top_up_by_bank_transfer_charge",
-                    "top_up_by_card_charge",
-                    "top_up_by_cash_or_cheque",
-                    "top_up_failed",
-                    "top_up_limits",
-                    "top_up_reverted",
-                    "topping_up_by_card",
-                    "transaction_charged_twice",
-                    "transfer_fee_charged",
-                    "transfer_into_account",
-                    "transfer_not_received_by_recipient",
-                    "transfer_timing",
-                    "unable_to_verify_identity",
-                    "verify_my_identity",
-                    "verify_source_of_funds",
-                    "verify_top_up",
-                    "virtual_card_not_working",
-                    "visa_or_mastercard",
-                    "why_verify_identity",
-                    "wrong_amount_of_cash_received",
-                    "wrong_exchange_rate_for_cash_withdrawal"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "banking_77",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 3373,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 376765,
-                "num_examples": 5000,
-                "dataset_name": null
-            }
-        },
-        "download_size": 214821,
-        "dataset_size": 380138,
-        "size_in_bytes": 594959
-    },
-    "terms_of_service": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Sentence": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "not potentially unfair",
-                    "potentially unfair"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "terms_of_service",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 10948,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 961820,
-                "num_examples": 5000,
-                "dataset_name": null
-            }
-        },
-        "download_size": 541547,
-        "dataset_size": 972768,
-        "size_in_bytes": 1514315
-    },
-    "tai_safety_research": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Title": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Abstract Note": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Url": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Publication Year": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Item Type": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Author": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Publication Title": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "TAI safety research",
-                    "not TAI safety research"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "tai_safety_research",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 54910,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 1634876,
-                "num_examples": 1639,
-                "dataset_name": null
-            }
-        },
-        "download_size": 948201,
-        "dataset_size": 1689786,
-        "size_in_bytes": 2637987
-    },
-    "neurips_impact_statement_risks": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Paper title": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Paper link": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Impact statement": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "doesn't mention a harmful application",
-                    "mentions a harmful application"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "neurips_impact_statement_risks",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 69037,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 198699,
-                "num_examples": 150,
-                "dataset_name": null
-            }
-        },
-        "download_size": 163355,
-        "dataset_size": 267736,
-        "size_in_bytes": 431091
-    },
-    "overruling": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Sentence": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "not overruling",
-                    "overruling"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "overruling",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 7424,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 431790,
-                "num_examples": 2350,
-                "dataset_name": null
-            }
-        },
-        "download_size": 277926,
-        "dataset_size": 439214,
-        "size_in_bytes": 717140
-    },
-    "systematic_review_inclusion": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Title": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Abstract": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Authors": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Journal": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "included",
-                    "not included"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "systematic_review_inclusion",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 52677,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 2344244,
-                "num_examples": 2243,
-                "dataset_name": null
-            }
-        },
-        "download_size": 1357407,
-        "dataset_size": 2396921,
-        "size_in_bytes": 3754328
-    },
-    "one_stop_english": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Article": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "advanced",
-                    "elementary",
-                    "intermediate"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "one_stop_english",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 201444,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 2090429,
-                "num_examples": 516,
-                "dataset_name": null
-            }
-        },
-        "download_size": 1409192,
-        "dataset_size": 2291873,
-        "size_in_bytes": 3701065
-    },
-    "tweet_eval_hate": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Tweet": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "hate speech",
-                    "not hate speech"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "tweet_eval_hate",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 7488,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 440048,
-                "num_examples": 2966,
-                "dataset_name": null
-            }
-        },
-        "download_size": 300542,
-        "dataset_size": 447536,
-        "size_in_bytes": 748078
-    },
-    "twitter_complaints": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Tweet text": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "complaint",
-                    "no complaint"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "twitter_complaints",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 5348,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 369564,
-                "num_examples": 3399,
-                "dataset_name": null
-            }
-        },
-        "download_size": 270136,
-        "dataset_size": 374912,
-        "size_in_bytes": 645048
-    },
-    "semiconductor_org_types": {
-        "description": "Large pre-trained language models have shown promise for few-shot learning, completing text-based tasks given only a few task-specific examples. Will models soon solve classification tasks that have so far been reserved for human research assistants? \n\n[RAFT](https://raft.elicit.org) is a few-shot classification benchmark that tests language models:\n\n- across multiple domains (lit review, tweets, customer interaction, etc.)\n- on economically valuable classification tasks (someone inherently cares about the task)\n- in a setting that mirrors deployment (50 examples per task, info retrieval allowed, hidden test set)\n",
-        "citation": "@InProceedings{huggingface:dataset,\ntitle = {A great new dataset},\nauthor={huggingface, Inc.\n},\nyear={2020}\n}\n",
-        "homepage": "https://raft.elicit.org",
-        "license": "",
-        "features": {
-            "Paper title": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "Organization name": {
-                "dtype": "string",
-                "_type": "Value"
-            },
-            "ID": {
-                "dtype": "int32",
-                "_type": "Value"
-            },
-            "Label": {
-                "names": [
-                    "Unlabeled",
-                    "company",
-                    "research institute",
-                    "university"
-                ],
-                "_type": "ClassLabel"
-            }
-        },
-        "builder_name": "parquet",
-        "dataset_name": "raft",
-        "config_name": "semiconductor_org_types",
-        "version": {
-            "version_str": "1.1.0",
-            "major": 1,
-            "minor": 1,
-            "patch": 0
-        },
-        "splits": {
-            "train": {
-                "name": "train",
-                "num_bytes": 8243,
-                "num_examples": 50,
-                "dataset_name": null
-            },
-            "test": {
-                "name": "test",
-                "num_bytes": 73864,
-                "num_examples": 449,
-                "dataset_name": null
-            }
-        },
-        "download_size": 53204,
-        "dataset_size": 82107,
-        "size_in_bytes": 135311
-    }
-}