Technology company running various social media platforms Meta Platforms Inc. (FB) announced having opened access to a large language model for artificial intelligence (AI) research purposes this week.
Large language models are natural languages processing systems which are trained on huge volume of texts and are capable of answering reading comprehension questions or generating new text.
According to the company, its model is the first 175-billion-parameter language model to be made available to the broader AI research community.
Through a blog post, Meta says the release of its Open Pretrained Transformer (OPT-175B) will improve the abilities of researchers to understand how large language models work. According to Meta, restrictions on access to such models have been hindering progress on efforts to improve their robustness and mitigate known issues such as bias and toxicity.
Some researchers have been expressing concerns related to the harms which can be spread through large language models. Meta says it’s hoping to increase to increase the diversity of voices defining the ethical considerations of such technologies.
To maintain integrity and prevent the misuse of the large language model, Meta said it will be releasing the model under a noncommercial license to focus on research use cases.