尽管大型语言模型（LLMS）经常产生令人印象深刻的输出，但它们也无法推理和事实。我们着手研究这些限制如何影响LLM回答基于困难的问题的能力和理由。我们应用了与人类一致的GPT-3（指示程序）回答多项选择体检问题（USMLE和MEDMCQA）和医学研究问题（PubMedQA）。我们调查了思想链（逐步思考）提示，接地（通过搜索结果增强提示）和很少的射击（以问答的示例来准备问题）。对于USMLE问题的子集，医疗领域专家审查并注释了模型的推理。总体而言，GPT-3取得了最新的机器学习绩效的重大改进。我们观察到GPT-3通常是知识渊博的，并且可以理解医疗问题。 GPT-3当面对一个无法回答的问题时，仍将尝试回答，通常会导致偏见的预测分布。 LLM与人类绩效不相同，但我们的结果表明，与医疗问题解决的推理模式的出现。我们推测，缩放模型和数据，增强及时对齐方式以及允许更好地完成完成的上下文化将足以使LLMS在此类任务上达到人级的性能。

Large language models (LLMs) have demonstrated impressive capabilities in natural language understanding and generation, but the quality bar for medical and clinical applications is high. Today, attempts to assess models' clinical knowledge typically rely on automated evaluations on limited benchmarks. There is no standard to evaluate model predictions and reasoning across a breadth of tasks. To address this, we present MultiMedQA, a benchmark combining six existing open question answering datasets spanning professional medical exams, research, and consumer queries; and HealthSearchQA, a new free-response dataset of medical questions searched online. We propose a framework for human evaluation of model answers along multiple axes including factuality, precision, possible harm, and bias. In addition, we evaluate PaLM (a 540-billion parameter LLM) and its instruction-tuned variant, Flan-PaLM, on MultiMedQA. Using a combination of prompting strategies, Flan-PaLM achieves state-of-the-art accuracy on every MultiMedQA multiple-choice dataset (MedQA, MedMCQA, PubMedQA, MMLU clinical topics), including 67.6% accuracy on MedQA (US Medical License Exam questions), surpassing prior state-of-the-art by over 17%. However, human evaluation reveals key gaps in Flan-PaLM responses. To resolve this we introduce instruction prompt tuning, a parameter-efficient approach for aligning LLMs to new domains using a few exemplars. The resulting model, Med-PaLM, performs encouragingly, but remains inferior to clinicians. We show that comprehension, recall of knowledge, and medical reasoning improve with model scale and instruction prompt tuning, suggesting the potential utility of LLMs in medicine. Our human evaluations reveal important limitations of today's models, reinforcing the importance of both evaluation frameworks and method development in creating safe, helpful LLM models for clinical applications.

最近的研究表明，理性或逐步思想链可用于改善多步推理任务的性能。我们重新考虑了理由的提示，提示了几次射击中的内部学习学习，其中（输入 - >输出）提示将扩展到（输入，理由 - >输出）提示。对于以理由为提示的提示，我们证明了现有的方法（依赖手动及时工程）如何受到可能损害绩效的次级理由。为了减轻这种脆弱性，我们提出了一个统一的授权合奏的统一框架，在该框架中，我们将输出空间中的理由抽样确定为可鲁棒提高性能的关键组成部分。该框架是一般的，可以轻松地扩展到常见的自然语言处理任务，即使传统上不利于中间步骤的任务，例如问题回答，单词感官歧义和情感分析。我们证明，与现有的提示方法相比，以理由为原理的合奏获得了更准确和可解释的结果 - 包括标准提示，没有理由和基于理由的链链链，同时通过相关理性同时提高了模型预测的解释性。

Recent work has shown that fine-tuning large pre-trained language models on a collection of tasks described via instructions, a.k.a. instruction-tuning, improves their zero and few-shot generalization to unseen tasks. However, there is a limited understanding of the performance trade-offs of different decisions made during the instruction-tuning process. These decisions include the scale and diversity of the instruction-tuning benchmark, different task sampling strategies, fine-tuning with and without demonstrations, training using specialized datasets for reasoning and dialogue, and finally, the fine-tuning objectives themselves. In this paper, we characterize the effect of instruction-tuning decisions on downstream task performance when scaling both model and benchmark sizes. To this end, we create OPT-IML Bench: a large benchmark for Instruction Meta-Learning (IML) of 2000 NLP tasks consolidated into task categories from 8 existing benchmarks, and prepare an evaluation framework to measure three types of model generalizations: to tasks from fully held-out categories, to held-out tasks from seen categories, and to held-out instances from seen tasks. Through the lens of this framework, we first present insights about instruction-tuning decisions as applied to OPT-30B and further exploit these insights to train OPT-IML 30B and 175B, which are instruction-tuned versions of OPT. OPT-IML demonstrates all three generalization abilities at both scales on four different evaluation benchmarks with diverse tasks and input formats -- PromptSource, FLAN, Super-NaturalInstructions, and UnifiedSKG. Not only does it significantly outperform OPT on all benchmarks but is also highly competitive with existing models fine-tuned on each specific benchmark. We release OPT-IML at both scales, together with the OPT-IML Bench evaluation framework.

我们介绍了Sparrow，这是一个寻求信息的对话代理，与提示的语言模型基线相比，训练有素，更有帮助，正确和无害。我们使用从人类反馈中的强化学习来培训我们的模型，以帮助人类评估者判断代理人的行为。首先，为了使我们的代理人更有帮助和无害，我们将良好对话的要求分解为代理人应遵循的自然语言规则，并分别向评估者询问每个规则。我们证明，这种崩溃使我们能够收集对代理行为的更多针对性的人类判断，并允许更有效的规则条件奖励模型。其次，我们的代理商在收集对模型声明的偏好判决时提供了支持事实主张的来源的证据。对于事实问题，麻雀提供的证据支持了78％的时间。比基线比基线更享受麻雀，同时对人类的对抗性探测更具弹性，在探测时只有8％的时间违反了我们的规则。最后，我们进行了广泛的分析，表明尽管我们的模型学会遵守我们的规则，但它可以表现出分布偏见。

预处理的大语言模型（LLM）广泛用于自然语言处理（NLP）的许多子场，通常被称为具有特定任务示例的优秀少数学习者。值得注意的是，思想链（COT）提示，这是一种通过分步答案示例引发复杂的多步推理的技术，在算术和符号推理中实现了最新的表演，难以置信的System-2任务不遵循LLMS的标准缩放定律。尽管这些成功通常归因于LLM的几次学习能力，但我们表明，LLM是通过在每个答案之前简单地添加“让我们逐步思考”而成为不错的零射击推理者。实验结果表明，使用相同的单个提示模板，我们的零射击功能明显优于零摄像机LLM在不同的基准推理任务上的零摄像机表现，包括算术（Multiarith，GSM8K，Aqua-Rat，SVAMP，SVAMP），符号推理（最后一个字母，字母，字母，字母，，，，，字母，字母）（最后一个字母），硬币翻转）和其他逻辑推理任务（日期理解，跟踪洗牌对象），而没有任何手工制作的几个示例，例如通过175B参数指令gpt模型将Multiarith的准确性从17.7％提高到78.7％，GSM8K从10.4％提高到40.7％，以及另一种现成的大型模型，540B参数Palm Palm的相似改进。在非常多样化的推理任务中，这个单一提示的多功能性暗示了LLM的尚未开发和研究的基本零拍功能，这表明可以通过简单提示来提取高级，多任务的广泛认知能力。我们希望我们的工作不仅可以作为具有挑战性的推理基准的最小零击基线，而且还强调了仔细探索和分析LLM中隐藏在LLM中的巨大的零拍知识的重要性，然后在制作Finetunning数据集或少数拍摄的典范之前。

最近已被证明大型语言模型在各种任务集中获得合理的零射普通化（Brown等，2020）。它已经假设这是语言模型的隐式多任务学习的结果，在语言模型中的预押（Radford等，2019）。可以通过明确的多任务学习直接引起零拍常规化？为了以缩放测试这个问题，我们开发一个系统，以便轻松地将任何自然语言任务映射到人类可读的提示表单中。我们转换一组大量的监督数据集，每个数据集都有多个提示，具有不同的措辞。这些提示的数据集允许基准测试模型执行完全看不见的任务的能力。我们介绍了一个普拉克尔编码器 - 解码器模型（Raffel等，2020; Lester等，2021），覆盖各种任务。该模型在多个标准数据集中达到强大的零点性能，通常优于其尺寸的型号超过16倍。此外，我们的方法对来自Big-替补基准测试的任务子集具有强烈性能，优于其尺寸的6倍。所有提示和培训的型号都可以在https://github.com/ bigscience-workshop / protectsource / httpsource / https：//huggingface.co/bigscience/t0pp。

We demonstrate that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even becoming competitive with prior state-ofthe-art fine-tuning approaches. Specifically, we train GPT-3, an autoregressive language model with 175 billion parameters, 10x more than any previous nonsparse 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. We also identify some datasets where GPT-3's few-shot learning still struggles, as well as some datasets where GPT-3 faces methodological issues related to training on large web corpora.

Large language models that are capable of zero or few-shot prompting approaches have given rise to the new research area of prompt engineering. Recent advances showed that for example Chain-of-Thought (CoT) prompts can improve arithmetic or common sense tasks significantly. We explore how such approaches fair with legal reasoning tasks and take the COLIEE entailment task based on the Japanese Bar exam for testing zero-shot/few-shot and fine-tuning approaches. Our findings show that while CoT prompting and fine-tuning with explanations approaches show improvements, the best results are produced by prompts that are derived from specific legal reasoning techniques such as IRAC (Issue, Rule, Application, Conclusion). Based on our experiments we improve the 2021 best result from 0.7037 accuracy to 0.8148 accuracy and beat the 2022 best system of 0.6789 accuracy with an accuracy of 0.7431.

The release of ChatGPT, a language model capable of generating text that appears human-like and authentic, has gained significant attention beyond the research community. We expect that the convincing performance of ChatGPT incentivizes users to apply it to a variety of downstream tasks, including prompting the model to simplify their own medical reports. To investigate this phenomenon, we conducted an exploratory case study. In a questionnaire, we asked 15 radiologists to assess the quality of radiology reports simplified by ChatGPT. Most radiologists agreed that the simplified reports were factually correct, complete, and not potentially harmful to the patient. Nevertheless, instances of incorrect statements, missed key medical findings, and potentially harmful passages were reported. While further studies are needed, the initial insights of this study indicate a great potential in using large language models like ChatGPT to improve patient-centered care in radiology and other medical domains.

推理是人类认知和智力的关键支柱。在过去的十年中，我们目睹了自然语言处理的巨大收益和大型语言模型的前所未有的缩放。最近的工作表征了很少射击技术的能力，例如思想链，可以在大语言模型中模仿人类的推理。这个标志性的功能很少，连同不断扩展的语言模型相结合，打开了解决各种任务的可能性的远景，例如数学单词问题，代码完成和常识性推理。促使思想链（COT）通过提供中间步骤并敦促模型遵循相同的过程，从而进一步推动了模型的性能。尽管具有令人信服的性能，但在这些模型中推理能力的起源却很少探索。这项工作启动了对大语言模型中推理机制的更深入了解的初步步骤。我们的工作围绕查询模型，同时在提示中控制除一个组件以外的所有组件外：符号，模式和文本。然后，我们分析查询之间的性能差异。我们的结果表明，在提示中存在事实模式对于COT的成功并不是必需的。尽管如此，我们从经验上表明，仅依靠模式也不足以获得高质量的结果。我们认为文本具有常识性知识和意义。我们详尽的经验分析提供了定性的例子，说明了文本和模式之间的共生关系。这种对COT的系统理解使我们能够设计简洁的思想链，被称为CCOT，在其中修剪文本和模式只能保留其关键角色，同时以PAR或更高的求解任务率交付。

Recent work pre-training Transformers with self-supervised objectives on large text corpora has shown great success when fine-tuned on downstream NLP tasks including text summarization. However, pre-training objectives tailored for abstractive text summarization have not been explored. Furthermore there is a lack of systematic evaluation across diverse domains. In this work, we propose pre-training large Transformer-based encoder-decoder models on massive text corpora with a new selfsupervised objective. In PEGASUS, important sentences are removed/masked from an input document and are generated together as one output sequence from the remaining sentences, similar to an extractive summary. We evaluated our best PEGASUS model on 12 downstream summarization tasks spanning news, science, stories, instructions, emails, patents, and legislative bills. Experiments demonstrate it achieves state-of-the-art performance on all 12 downstream datasets measured by ROUGE scores. Our model also shows surprising performance on low-resource summarization, surpassing previous state-of-the-art results on 6 datasets with only 1000 examples. Finally we validated our results using human evaluation and show that our model summaries achieve human performance on multiple datasets.

我们探索如何产生一系列思想 - 一系列中间推理步骤 - 显着提高了大语言模型执行复杂推理的能力。特别是，我们通过一种称为“思想链”提示的简单方法在足够大的语言模型中自然出现这种推理能力，在此过程中，一些思想示范被作为提示的示例提供了。三种大语模型的实验表明，促使思想链提高了一系列算术，常识和象征性推理任务的性能。经验收益可能会引人注目。例如，仅使用八个思想范围的540B参数语言模型才能在数学单词问题的GSM8K基准上实现最新的精度，甚至超过了带有验证器的Fineted GPT-3。

Generating a chain of thought (CoT) can increase large language model (LLM) performance on a wide range of tasks. Zero-shot CoT evaluations, however, have been conducted primarily on logical tasks (e.g. arithmetic, commonsense QA). In this paper, we perform a controlled evaluation of zero-shot CoT across two sensitive domains: harmful questions and stereotype benchmarks. We find that using zero-shot CoT reasoning in a prompt can significantly increase a model's likelihood to produce undesirable output. Without future advances in alignment or explicit mitigation instructions, zero-shot CoT should be avoided on tasks where models can make inferences about marginalized groups or harmful topics.

Finetuning language models on a collection of datasets phrased as instructions has been shown to improve model performance and generalization to unseen tasks. In this paper we explore instruction finetuning with a particular focus on (1) scaling the number of tasks, (2) scaling the model size, and (3) finetuning on chain-of-thought data. We find that instruction finetuning with the above aspects dramatically improves performance on a variety of model classes (PaLM, T5, U-PaLM), prompting setups (zero-shot, few-shot, CoT), and evaluation benchmarks (MMLU, BBH, TyDiQA, MGSM, open-ended generation). For instance, Flan-PaLM 540B instruction-finetuned on 1.8K tasks outperforms PALM 540B by a large margin (+9.4% on average). Flan-PaLM 540B achieves state-of-the-art performance on several benchmarks, such as 75.2% on five-shot MMLU. We also publicly release Flan-T5 checkpoints, which achieve strong few-shot performance even compared to much larger models, such as PaLM 62B. Overall, instruction finetuning is a general method for improving the performance and usability of pretrained language models.

语言模型可以根据给定的文化背景产生有害和偏置的输出并表现出不良行为。我们提出了一种将语言模型适应社会（PALM）与值目标数据集的过程，以通过在反映预定的一组目标值集合的数据集上进行制备和微调来显着地改变模型行为的迭代过程。我们使用三个指标评估我们的进程：具有人类评估的定量指标，将输出遵守目标值，毒性评分对产出;和定性度量分析与给定社会类别相关的最常见的单词。通过每次迭代，我们根据来自评估的观察到的缺点添加其他培训数据集示例。与基线和控制模型相比，PALMS在所有指标上显着更好地为广泛的GPT-3语言模型尺寸进行了基线和控制模型，而不会影响能力完整性。我们发现PALMS的有效性随模型规模而增加。我们表明，显着调整语言模型行为与小型手腕策划数据集是可行的。

Developing safe and useful general-purpose AI systems will require us to make progress on scalable oversight: the problem of supervising systems that potentially outperform us on most skills relevant to the task at hand. Empirical work on this problem is not straightforward, since we do not yet have systems that broadly exceed our abilities. This paper discusses one of the major ways we think about this problem, with a focus on how to turn it into one that can be productively studied empirically. We first present an experimental design centered on choosing tasks for which human specialists succeed but unaided humans and current general AI systems fail. We then present a proof-of-concept experiment following meant to demonstrate a key feature of this experimental design and show its viability with two question-answering tasks: MMLU and time-limited QuALITY. On these tasks, we find that human participants who interact with an unreliable large-language-model dialog assistant through chat -- a trivial baseline strategy for scalable oversight -- substantially outperform both the model alone and their own unaided performance. These results are an encouraging sign that scalable oversight will be tractable to study with present models and bolster recent findings that large language models can productively assist humans with difficult tasks.

State-of-the-art computer vision systems are trained to predict a fixed set of predetermined object categories. This restricted form of supervision limits their generality and usability since additional labeled data is needed to specify any other visual concept. Learning directly from raw text about images is a promising alternative which leverages a much broader source of supervision. We demonstrate that the simple pre-training task of predicting which caption goes with which image is an efficient and scalable way to learn SOTA image representations from scratch on a dataset of 400 million (image, text) pairs collected from the internet. After pre-training, natural language is used to reference learned visual concepts (or describe new ones) enabling zero-shot transfer of the model to downstream tasks. We study the performance of this approach by benchmarking on over 30 different existing computer vision datasets, spanning tasks such as OCR, action recognition in videos, geo-localization, and many types of fine-grained object classification. The model transfers non-trivially to most tasks and is often competitive with a fully supervised baseline without the need for any dataset specific training. For instance, we match the accuracy of the original ResNet-50 on ImageNet zero-shot without needing to use any of the 1.28 million training examples it was trained on. We release our code and pre-trained model weights at https://github.com/OpenAI/CLIP.

大型语言模型在各种任务上显示出令人印象深刻的几次结果。但是，当知识是此类结果的关键时，就像问题回答和事实检查之类的任务一样，似乎需要存储知识的大量参数计数。众所周知，检索增强模型可以在不需要多个参数的情况下在知识密集的任务上表现出色，但是目前尚不清楚它们是否在几个弹药设置中工作。在这项工作中，我们介绍了地图集，这是一个经过精心设计和预先训练的增强语言模型，能够通过很少的培训示例学习知识密集型任务。我们对包括MMLU，苏格兰短裙和归类等各种任务进行评估，并研究文档索引内容的影响，表明它可以很容易地进行更新。值得注意的是，在自然问题上仅使用64个示例在自然问题上达到超过42 \％的准确性，尽管参数少了50倍，但比540B参数模型的表现优于540b参数模型。

A long-running goal of the clinical NLP community is the extraction of important variables trapped in clinical notes. However, roadblocks have included dataset shift from the general domain and a lack of public clinical corpora and annotations. In this work, we show that large language models, such as InstructGPT, perform well at zero- and few-shot information extraction from clinical text despite not being trained specifically for the clinical domain. Whereas text classification and generation performance have already been studied extensively in such models, here we additionally demonstrate how to leverage them to tackle a diverse set of NLP tasks which require more structured outputs, including span identification, token-level sequence classification, and relation extraction. Further, due to the dearth of available data to evaluate these systems, we introduce new datasets for benchmarking few-shot clinical information extraction based on a manual re-annotation of the CASI dataset for new tasks. On the clinical extraction tasks we studied, the GPT-3 systems significantly outperform existing zero- and few-shot baselines.