{"mid_word_1_anagrams": {"description": "Unscramble is a small battery of 5 \u201ccharacter manipulation\u201d tasks. Each task\ninvolves giving the model a word distorted by some combination of scrambling,\naddition, or deletion of characters, and asking it to recover the original word.\n", "citation": "@inproceedings{NEURIPS2020_1457c0d6,\n    author = {Brown, Tom and Mann, Benjamin and Ryder, Nick and Subbiah, Melanie and Kaplan, Jared D and Dhariwal, Prafulla and Neelakantan, Arvind and Shyam, Pranav and Sastry, Girish and Askell, Amanda and Agarwal, Sandhini and Herbert-Voss, Ariel and Krueger, Gretchen and Henighan, Tom and Child, Rewon and Ramesh, Aditya and Ziegler, Daniel and Wu, Jeffrey and Winter, Clemens and Hesse, Chris and Chen, Mark and Sigler, Eric and Litwin, Mateusz and Gray, Scott and Chess, Benjamin and Clark, Jack and Berner, Christopher and McCandlish, Sam and Radford, Alec and Sutskever, Ilya and Amodei, Dario},\n    booktitle = {Advances in Neural Information Processing Systems},\n    editor = {H. Larochelle and M. Ranzato and R. Hadsell and M. F. Balcan and H. Lin},\n    pages = {1877--1901},\n    publisher = {Curran Associates, Inc.},\n    title = {Language Models are Few-Shot Learners},\n    url = {https://proceedings.neurips.cc/paper/2020/file/1457c0d6bfcb4967418bfb8ac142f64a-Paper.pdf},\n    volume = {33},\n    year = {2020}\n}\n", "homepage": "https://github.com/openai/gpt-3/tree/master/data", "license": "", "features": {"context": {"dtype": "string", "id": null, "_type": "Value"}, "completion": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "unscramble", "config_name": "mid_word_1_anagrams", "version": {"version_str": "0.0.1", "description": null, "major": 0, "minor": 0, "patch": 1}, "splits": {"validation": {"name": "validation", "num_bytes": 271516, "num_examples": 10000, "dataset_name": "unscramble"}}, "download_checksums": {"https://raw.githubusercontent.com/openai/gpt-3/master/data/mid_word_1_anagrams.jsonl.gz": {"num_bytes": 106533, "checksum": "6768a86896083199de4815d4964cb2f6f1046476cfd80c2a562784f182905979"}}, "download_size": 106533, "post_processing_size": null, "dataset_size": 271516, "size_in_bytes": 378049}, "mid_word_2_anagrams": {"description": "Unscramble is a small battery of 5 \u201ccharacter manipulation\u201d tasks. Each task\ninvolves giving the model a word distorted by some combination of scrambling,\naddition, or deletion of characters, and asking it to recover the original word.\n", "citation": "@inproceedings{NEURIPS2020_1457c0d6,\n    author = {Brown, Tom and Mann, Benjamin and Ryder, Nick and Subbiah, Melanie and Kaplan, Jared D and Dhariwal, Prafulla and Neelakantan, Arvind and Shyam, Pranav and Sastry, Girish and Askell, Amanda and Agarwal, Sandhini and Herbert-Voss, Ariel and Krueger, Gretchen and Henighan, Tom and Child, Rewon and Ramesh, Aditya and Ziegler, Daniel and Wu, Jeffrey and Winter, Clemens and Hesse, Chris and Chen, Mark and Sigler, Eric and Litwin, Mateusz and Gray, Scott and Chess, Benjamin and Clark, Jack and Berner, Christopher and McCandlish, Sam and Radford, Alec and Sutskever, Ilya and Amodei, Dario},\n    booktitle = {Advances in Neural Information Processing Systems},\n    editor = {H. Larochelle and M. Ranzato and R. Hadsell and M. F. Balcan and H. Lin},\n    pages = {1877--1901},\n    publisher = {Curran Associates, Inc.},\n    title = {Language Models are Few-Shot Learners},\n    url = {https://proceedings.neurips.cc/paper/2020/file/1457c0d6bfcb4967418bfb8ac142f64a-Paper.pdf},\n    volume = {33},\n    year = {2020}\n}\n", "homepage": "https://github.com/openai/gpt-3/tree/master/data", "license": "", "features": {"context": {"dtype": "string", "id": null, "_type": "Value"}, "completion": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "unscramble", "config_name": "mid_word_2_anagrams", "version": {"version_str": "0.0.1", "description": null, "major": 0, "minor": 0, "patch": 1}, "splits": {"validation": {"name": "validation", "num_bytes": 282654, "num_examples": 10000, "dataset_name": "unscramble"}}, "download_checksums": {"https://raw.githubusercontent.com/openai/gpt-3/master/data/mid_word_2_anagrams.jsonl.gz": {"num_bytes": 109091, "checksum": "c3d839d09a7954b78a27cd2cd75d4ed0488656c56ef4dbd741a005343826cb01"}}, "download_size": 109091, "post_processing_size": null, "dataset_size": 282654, "size_in_bytes": 391745}, "cycle_letters_in_word": {"description": "Unscramble is a small battery of 5 \u201ccharacter manipulation\u201d tasks. Each task\ninvolves giving the model a word distorted by some combination of scrambling,\naddition, or deletion of characters, and asking it to recover the original word.\n", "citation": "@inproceedings{NEURIPS2020_1457c0d6,\n    author = {Brown, Tom and Mann, Benjamin and Ryder, Nick and Subbiah, Melanie and Kaplan, Jared D and Dhariwal, Prafulla and Neelakantan, Arvind and Shyam, Pranav and Sastry, Girish and Askell, Amanda and Agarwal, Sandhini and Herbert-Voss, Ariel and Krueger, Gretchen and Henighan, Tom and Child, Rewon and Ramesh, Aditya and Ziegler, Daniel and Wu, Jeffrey and Winter, Clemens and Hesse, Chris and Chen, Mark and Sigler, Eric and Litwin, Mateusz and Gray, Scott and Chess, Benjamin and Clark, Jack and Berner, Christopher and McCandlish, Sam and Radford, Alec and Sutskever, Ilya and Amodei, Dario},\n    booktitle = {Advances in Neural Information Processing Systems},\n    editor = {H. Larochelle and M. Ranzato and R. Hadsell and M. F. Balcan and H. Lin},\n    pages = {1877--1901},\n    publisher = {Curran Associates, Inc.},\n    title = {Language Models are Few-Shot Learners},\n    url = {https://proceedings.neurips.cc/paper/2020/file/1457c0d6bfcb4967418bfb8ac142f64a-Paper.pdf},\n    volume = {33},\n    year = {2020}\n}\n", "homepage": "https://github.com/openai/gpt-3/tree/master/data", "license": "", "features": {"context": {"dtype": "string", "id": null, "_type": "Value"}, "completion": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "unscramble", "config_name": "cycle_letters_in_word", "version": {"version_str": "0.0.1", "description": null, "major": 0, "minor": 0, "patch": 1}, "splits": {"validation": {"name": "validation", "num_bytes": 282654, "num_examples": 10000, "dataset_name": "unscramble"}}, "download_checksums": {"https://raw.githubusercontent.com/openai/gpt-3/master/data/cycle_letters_in_word.jsonl.gz": {"num_bytes": 98451, "checksum": "1689c9002bb8c5988bf5f05e977c9db92f57932c1b5a38998c29ac0dd71e1d42"}}, "download_size": 98451, "post_processing_size": null, "dataset_size": 282654, "size_in_bytes": 381105}, "random_insertion_in_word": {"description": "Unscramble is a small battery of 5 \u201ccharacter manipulation\u201d tasks. Each task\ninvolves giving the model a word distorted by some combination of scrambling,\naddition, or deletion of characters, and asking it to recover the original word.\n", "citation": "@inproceedings{NEURIPS2020_1457c0d6,\n    author = {Brown, Tom and Mann, Benjamin and Ryder, Nick and Subbiah, Melanie and Kaplan, Jared D and Dhariwal, Prafulla and Neelakantan, Arvind and Shyam, Pranav and Sastry, Girish and Askell, Amanda and Agarwal, Sandhini and Herbert-Voss, Ariel and Krueger, Gretchen and Henighan, Tom and Child, Rewon and Ramesh, Aditya and Ziegler, Daniel and Wu, Jeffrey and Winter, Clemens and Hesse, Chris and Chen, Mark and Sigler, Eric and Litwin, Mateusz and Gray, Scott and Chess, Benjamin and Clark, Jack and Berner, Christopher and McCandlish, Sam and Radford, Alec and Sutskever, Ilya and Amodei, Dario},\n    booktitle = {Advances in Neural Information Processing Systems},\n    editor = {H. Larochelle and M. Ranzato and R. Hadsell and M. F. Balcan and H. Lin},\n    pages = {1877--1901},\n    publisher = {Curran Associates, Inc.},\n    title = {Language Models are Few-Shot Learners},\n    url = {https://proceedings.neurips.cc/paper/2020/file/1457c0d6bfcb4967418bfb8ac142f64a-Paper.pdf},\n    volume = {33},\n    year = {2020}\n}\n", "homepage": "https://github.com/openai/gpt-3/tree/master/data", "license": "", "features": {"context": {"dtype": "string", "id": null, "_type": "Value"}, "completion": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "unscramble", "config_name": "random_insertion_in_word", "version": {"version_str": "0.0.1", "description": null, "major": 0, "minor": 0, "patch": 1}, "splits": {"validation": {"name": "validation", "num_bytes": 353981, "num_examples": 10000, "dataset_name": "unscramble"}}, "download_checksums": {"https://raw.githubusercontent.com/openai/gpt-3/master/data/random_insertion_in_word.jsonl.gz": {"num_bytes": 143626, "checksum": "72e65d83da53d15752ee0c47379509de149ddbad32d61184e5991df29616b78a"}}, "download_size": 143626, "post_processing_size": null, "dataset_size": 353981, "size_in_bytes": 497607}, "reversed_words": {"description": "Unscramble is a small battery of 5 \u201ccharacter manipulation\u201d tasks. Each task\ninvolves giving the model a word distorted by some combination of scrambling,\naddition, or deletion of characters, and asking it to recover the original word.\n", "citation": "@inproceedings{NEURIPS2020_1457c0d6,\n    author = {Brown, Tom and Mann, Benjamin and Ryder, Nick and Subbiah, Melanie and Kaplan, Jared D and Dhariwal, Prafulla and Neelakantan, Arvind and Shyam, Pranav and Sastry, Girish and Askell, Amanda and Agarwal, Sandhini and Herbert-Voss, Ariel and Krueger, Gretchen and Henighan, Tom and Child, Rewon and Ramesh, Aditya and Ziegler, Daniel and Wu, Jeffrey and Winter, Clemens and Hesse, Chris and Chen, Mark and Sigler, Eric and Litwin, Mateusz and Gray, Scott and Chess, Benjamin and Clark, Jack and Berner, Christopher and McCandlish, Sam and Radford, Alec and Sutskever, Ilya and Amodei, Dario},\n    booktitle = {Advances in Neural Information Processing Systems},\n    editor = {H. Larochelle and M. Ranzato and R. Hadsell and M. F. Balcan and H. Lin},\n    pages = {1877--1901},\n    publisher = {Curran Associates, Inc.},\n    title = {Language Models are Few-Shot Learners},\n    url = {https://proceedings.neurips.cc/paper/2020/file/1457c0d6bfcb4967418bfb8ac142f64a-Paper.pdf},\n    volume = {33},\n    year = {2020}\n}\n", "homepage": "https://github.com/openai/gpt-3/tree/master/data", "license": "", "features": {"context": {"dtype": "string", "id": null, "_type": "Value"}, "completion": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "unscramble", "config_name": "reversed_words", "version": {"version_str": "0.0.1", "description": null, "major": 0, "minor": 0, "patch": 1}, "splits": {"validation": {"name": "validation", "num_bytes": 282654, "num_examples": 10000, "dataset_name": "unscramble"}}, "download_checksums": {"https://raw.githubusercontent.com/openai/gpt-3/master/data/reversed_words.jsonl.gz": {"num_bytes": 91917, "checksum": "133a08f875cd6c1ef8608a3233571a773881cc27b1c707de738cc6543439332a"}}, "download_size": 91917, "post_processing_size": null, "dataset_size": 282654, "size_in_bytes": 374571}}