Transcription of arXiv:2005.14165v4 [cs.CL] 22 Jul 2020
{{id}} {{{paragraph}}}
Example: biology
arXiv:2005.14165v4 [cs.CL] 22 Jul 2020
Language Models are Few-Shot LearnersTom B. Brown Benjamin Mann Nick Ryder Melanie Subbiah Jared Kaplan Prafulla DhariwalArvind NeelakantanPranav ShyamGirish SastryAmanda AskellSandhini AgarwalAriel Herbert-VossGretchen KruegerTom HenighanRewon ChildAditya RameshDaniel M. ZieglerJeffrey WuClemens WinterChristopher HesseMark ChenEric SiglerMateusz LitwinScott GrayBenjamin ChessJack ClarkChristopher BernerSam McCandlishAlec RadfordIlya SutskeverDario AmodeiOpenAIAbstractRecent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-trainingon a large corpus of text followed by fine-tuning on a specific task. While typically task-agnosticin architecture, this method still requires task-specific fine-tuning datasets of thousands or tens ofthousands of examples. By contrast, humans can generally perform a new language task from onlya few examples or from simple instructions something which current NLP systems still largelystruggle to do.
loop of meta-learning. We further specialize the description to “zero-shot”, “one-shot”, or “few-shot” depending on how many demonstrations are provided at inference time. These terms are intended to remain agnostic on the question of whether the model
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: