Add topic "GPT-3" Accepted
Changes: 5
-
Add Language Models are Few-Shot Learners
- Title
-
- Unchanged
- Language Models are Few-Shot Learners
- Type
-
- Unchanged
- Paper
- Created
-
- Unchanged
- 2020-01-01
- Description
-
- Unchanged
- Humans generally perform a new language task from only a few examples - something which current NLP systems still largely struggle to do. Here we show that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even reaching competitiveness with prior state-of-the-art fine-tuning approaches. Specifically, we train GPT-3, an autoregressive language model with 175 billion parameters, 10x more than any previous non-sparse language model, and test its performance in the few-shot setting. For all tasks, GPT-3 is applied without any gradient updates or fine-tuning, with tasks and few-shot demonstrations specified purely via text interaction with the model. GPT-3 achieves strong performance on many NLP datasets, including translation, question-answering, and cloze tasks, as well as several tasks that require on-the-fly reasoning or domain adaptation, such as unscrambling words, using a novel word in a sentence, or performing 3-digit arithmetic.
- Link
-
- Unchanged
- http://arxiv.org/abs/2005.14165
- Identifier
-
- Unchanged
- https://github.com/openai/gpt-3
Resource | v1 | current (v1) -
Add Tempering Expectations for GPT-3 and OpenAI’s API
- Title
-
- Unchanged
- Tempering Expectations for GPT-3 and OpenAI’s API
- Type
-
- Unchanged
- Web
- Created
-
- Unchanged
- 2020-01-01
- Description
-
- Unchanged
- GPT-3 is indeed a large step forward for AI text-generation, but there are very many caveats with the popular demos and use cases. This blog post discuses exactly those.
- Link
-
- Unchanged
- https://minimaxir.com/2020/07/gpt3-expectations/
- Identifier
-
- Unchanged
- no value
Resource | v1 | current (v1) -
Add GPT-3
- Title
-
- Unchanged
- GPT-3
- Description
-
- Unchanged
- GPT-3 is a transformer based text generation neural network released by OpenAI on May 29th 2020.
- Link
-
- Unchanged
- https://openai.com/blog/openai-api/
Topic | v1 | current (v1) -
Add GPT-3 has official Language Models are Few-Shot Learners
- Current
- has official
Topic to resource relation | v1 -
Add GPT-3 cons given in Tempering Expectations for GPT-3 and OpenAI’s API
- Current
- cons given in
Topic to resource relation | v1