We introduce Codex, a GPT language model fine-tuned on publicly available code from GitHub, and study its Python code-writing capabilities. A distinct production version of Codex powers GitHub Copilot. On HumanEval, a new evaluation set we release to measure functional correctness for synthesizing programs from docstrings, our model solves 28.8% of the problems, while GPT-3 solves 0% and GPT-J solves 11.4%. Furthermore, we find that repeated sampling from the model is a surprisingly effective strategy for producing working solutions to difficult prompts. Using this method, we solve 70.2% of our problems with 100 samples per problem. Careful investigation of our model reveals its limitations, including difficulty with docstrings describing long chains of operations and with binding operations to variables. Finally, we discuss the potential broader impacts of deploying powerful code generation technologies, covering safety, security, and economics.
Homepage: https://github.com/openai/human-eval
## Citation
```
@article{chen2021codex,
title={Evaluating Large Language Models Trained on Code},
author={Mark Chen and Jerry Tworek and Heewoo Jun and Qiming Yuan and Henrique Ponde de Oliveira Pinto and Jared Kaplan and Harri Edwards and Yuri Burda and Nicholas Joseph and Greg Brockman and Alex Ray and Raul Puri and Gretchen Krueger and Michael Petrov and Heidy Khlaaf and Girish Sastry and Pamela Mishkin and Brooke Chan and Scott Gray and Nick Ryder and Mikhail Pavlov and Alethea Power and Lukasz Kaiser and Mohammad Bavarian and Clemens Winter and Philippe Tillet and Felipe Petroski Such and Dave Cummings and Matthias Plappert and Fotios Chantzis and Elizabeth Barnes and Ariel Herbert-Voss and William Hebgen Guss and Alex Nichol and Alex Paino and Nikolas Tezak and Jie Tang and Igor Babuschkin and Suchir Balaji and Shantanu Jain and William Saunders and Christopher Hesse and Andrew N. Carr and Jan Leike and Josh Achiam and Vedant Misra and Evan Morikawa and Alec Radford and Matthew Knight and Miles Brundage and Mira Murati and Katie Mayer and Peter Welinder and Bob McGrew and Dario Amodei and Sam McCandlish and Ilya Sutskever and Wojciech Zaremba},
year={2021},
eprint={2107.03374},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```
### Groups and Tasks
#### Groups
* Not part of a group yet.
#### Tasks
-`humaneval` pass@1
-`humaneval_64` pass@64 variant
### Checklist
For adding novel benchmarks/datasets to the library:
* [ ] Is the task an existing benchmark in the literature?
* [ ] Have you referenced the original paper that introduced the task?
* [ ] If yes, does the original paper provide a reference implementation? If so, have you checked against the reference implementation and documented how to run such a test?
If other tasks on this dataset are already supported:
* [ ] Is the "Main" variant of this task clearly denoted?
* [ ] Have you provided a short sentence in a README on what each new variant adds / evaluates?
* [ ] Have you noted which, if any, published evaluation setups are matched by this variant?
),f"`nltk` version {nltk_version} is not >= {NLTK_MIN_VERSION}. Please update `nltk` before proceeding--older versions are vulnerable to a remote code execution vulnerability."
),f"`nltk` version {nltk_version} is not >= {NLTK_MIN_VERSION}. Please update `nltk` before proceeding--older versions are vulnerable to a remote code execution vulnerability."
The Japanese LLM Leaderboard evaluates language models based on a wide range of NLP tasks that reflect the characteristics of the Japanese language.
### Groups, Tags, and Tasks
#### Groups
*`japanese_leaderboard`: runs all tasks defined in this directory
#### Tasks
##### Generation Evaluation
*`ja_leaderboard_jaqket_v2`: The JAQKET dataset is designed for Japanese question answering research, featuring quiz-like questions with answers derived from Wikipedia article titles. [Source](https://github.com/kumapo/JAQKET-dataset)
*`ja_leaderboard_mgsm`: Multilingual Grade School Math Benchmark (MGSM) is a benchmark of grade-school math problems, proposed in the paper Language models are multilingual chain-of-thought reasoners. [Source](https://huggingface.co/datasets/juletxara/mgsm)
*`ja_leaderboard_xlsum`: This is the filtered Japanese subset of XL-Sum. [Source](https://github.com/csebuetnlp/xl-sum)
*`ja_leaderboard_jsquad`: JSQuAD is a Japanese version of SQuAD, a reading comprehension dataset. Each instance in the dataset consists of a question regarding a given context (Wikipedia article) and its answer. JSQuAD is based on SQuAD 1.1 (there are no unanswerable questions). [Source](https://github.com/yahoojapan/JGLUE)
##### Multi-Choice/Classification Evaluation
*`ja_leaderboard_jcommonsenseqa`: JCommonsenseQA is a Japanese version of CommonsenseQA, which is a multiple-choice question answering dataset that requires commonsense reasoning ability. [Source](https://github.com/yahoojapan/JGLUE)
*`ja_leaderboard_jnli`: JNLI is a Japanese version of the NLI (Natural Language Inference) dataset. The inference relations are entailment (含意), contradiction (矛盾), and neutral (中立). [Source](https://github.com/yahoojapan/JGLUE)
*`ja_leaderboard_marc_ja`: MARC-ja is a text classification dataset based on the Japanese portion of Multilingual Amazon Reviews Corpus (MARC). [Source](https://github.com/yahoojapan/JGLUE)
*`ja_leaderboard_xwinograd`: This is the Japanese portion of XWinograd. [Source](https://huggingface.co/datasets/polm-stability/xwinograd-ja)
### Citation
```bibtex
@inproceedings{ja_leaderboard_jaqket_v2,
title={JAQKET: クイズを題材にした日本語 QA データセットの構築},
author={鈴木正敏 and 鈴木潤 and 松田耕史 and ⻄田京介 and 井之上直也},
For adding novel benchmarks/datasets to the library:
* [x] Is the task an existing benchmark in the literature?
* [x] Have you referenced the original paper that introduced the task?
* [ ] If yes, does the original paper provide a reference implementation? If so, have you checked against the reference implementation and documented how to run such a test?
If other tasks on this dataset are already supported:
* [ ] Is the "Main" variant of this task clearly denoted?
* [ ] Have you provided a short sentence in a README on what each new variant adds / evaluates?
* [ ] Have you noted which, if any, published evaluation setups are matched by this variant?
returnf"`{name}` is required for `japanese_leaderboard`, please install `{name}` via pip install lm_eval[japanese_leaderboard] or pip install -e .[japanese_leaderboard]"
