From 953530fc24dea8670d2dcb09ef054b416da96346 Mon Sep 17 00:00:00 2001 From: Jeff Wu Date: Wed, 6 Mar 2019 11:30:53 -0800 Subject: [PATCH] update readme with usage caveats and calls for research MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This write-up was loosely inspired in part by Mitchell et al.’s work on [Model Cards for Model Reporting](https://arxiv.org/abs/1810.03993). Adding such model usage sections could be good practice in general for open source research projects with potentially broad applications. --- README.md | 20 ++++++++++++++++++-- 1 file changed, 18 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 3900326..8425e9a 100644 --- a/README.md +++ b/README.md @@ -6,6 +6,22 @@ For now, we have only released a smaller (117M parameter) version of GPT-2. See more details in our [blog post](https://blog.openai.com/better-language-models/). +## Usage + +This repository is meant to be a starting point for researchers and engineers to experiment with GPT-2-117M. While GPT-2-117M is less proficient than GPT-2-1.5B, it is useful for a wide range of research and applications which could also apply to larger models. + +### Some caveats + +- GPT-2-117M robustness and worst case behaviors are not well-understood. As with any machine-learned model, carefully evaluate GPT-2-117M for your use case, especially if used without fine-tuning or in safety-critical applications where reliability is important. +- The dataset our GPT-2-117M was trained on contains many texts with [biases](https://twitter.com/TomerUllman/status/1101485289720242177) and factual inaccuracies, and thus GPT-2-117M is likely to be biased and inaccurate as well. +- To avoid having samples mistaken as human-written, we recommend clearly labeling samples as synthetic before wide dissemination. Our models are often incoherent or inaccurate in subtle ways, which takes more than a quick read for a human to notice. + +### Work with us + +Please [let us know](mailto:languagequestions@openai.com) if you’re doing interesting research with or working on applications of GPT-2-117M! We’re especially interested in hearing from and potentially working with those who are studying +- Potential malicious use cases and defenses against them (e.g. the detectability of synthetic text) +- The extent of problematic content (e.g. bias) being baked into the models and effective mitigations + ## Installation Git clone this repository, and `cd` into directory for remaining commands @@ -53,7 +69,7 @@ and a valid install of [nvidia-docker 2.0](https://github.com/nvidia/nvidia-dock docker run --runtime=nvidia -it gpt-2 bash ``` -## Usage +## Sampling scripts | WARNING: Samples are unfiltered and may contain offensive content. | | --- | @@ -120,4 +136,4 @@ We are still considering release of the larger models. ## License -MIT +[MIT](./LICENSE)