Head of Safety Systems at OpenAI. Deep technical blog posts on AI research.
Special thanks to John Schulman for a lot of super valuable feedback and direct edits on this post. Test time compute (Graves et al. 2016, Ling, et al. 2017, Cobbe et al. 2021) and Chain-of-thought (CoT) (Wei et al. 2022, Nye et al. 2021), have led to significant improvements in model performance, while raising many research questions. This post aims to review recent developments in how to effectively use test-time compute (i.e. “thinking time”) and why it helps.
Reward hacking occurs when a reinforcement learning (RL) agent exploits flaws or ambiguities in the reward function to achieve high rewards, without genuinely learning or completing the intended task. Reward hacking exists because RL environments are often imperfect, and it is fundamentally challenging to accurately specify a reward function. With the rise of language models generalizing to a broad spectrum of tasks and RLHF becomes a de facto method for alignment training, reward hacking in RL
Hallucination in large language models usually refers to the model generating unfaithful, fabricated, inconsistent, or nonsensical content. As a term, hallucination has been somewhat generalized to cases when the model makes mistakes. Here, I would like to narrow down the problem of hallucination to cases where the model output is fabricated and not grounded by either the provided context or world knowledge. There are two types of hallucination: In-context hallucination: The model output should
Diffusion models have demonstrated strong results on image synthesis in past years. Now the research community has started working on a harder task—using it for video generation. The task itself is a superset of the image case, since an image is a video of 1 frame, and it is much more challenging because: It has extra requirements on temporal consistency across frames in time, which naturally demands more world knowledge to be encoded into the model. In comparison to text or images, it is
[Special thank you to Ian Kivlichan for many useful pointers (E.g. the 100+ year old Nature paper “Vox populi”) and nice feedback. 🙏 ] High-quality data is the fuel for modern data deep learning model training. Most of the task-specific labeled data comes from human annotation, such as classification task or RLHF labeling (which can be constructed as classification format) for LLM alignment training. Lots of ML techniques in the post can help with data quality, but fundamentally huma
The use of large language models in the real world has strongly accelerated by the launch of ChatGPT. We (including my team at OpenAI, shoutout to them) have invested a lot of effort to build default safe behavior into the model during the alignment process (e.g. via RLHF). However, adversarial attacks or jailbreak prompts could potentially trigger the model to output something undesired. A large body of ground work on adversarial attacks is on images, and differently it operates in the continuo
Building agents with LLM (large language model) as its core controller is a cool concept. Several proof-of-concepts demos, such as AutoGPT, GPT-Engineer and BabyAGI, serve as inspiring examples. The potentiality of LLM extends beyond generating well-written copies, stories, essays and programs; it can be framed as a powerful general problem solver. Agent System Overview In a LLM-powered autonomous agent system, LLM functions as the agent’s brain, complemented by several key components: Pl
Prompt Engineering, also known as In-Context Prompting, refers to methods for how to communicate with LLM to steer its behavior for desired outcomes without updating the model weights. It is an empirical science and the effect of prompt engineering methods can vary a lot among models, thus requiring heavy experimentation and heuristics. This post only focuses on prompt engineering for autoregressive language models, so nothing with Cloze tests, image generation or multimodality models. At its co
Many new Transformer architecture improvements have been proposed since my last post on “The Transformer Family” about three years ago. Here I did a big refactoring and enrichment of that 2020 post — restructure the hierarchy of sections and improve many sections with more recent papers. Version 2.0 is a superset of the old version, about twice the length. Notations Symbol Meaning $d$ The model size / hidden
[Updated on 2023-01-24: add a small section on Distillation.] Large transformer models are mainstream nowadays, creating SoTA results for a variety of tasks. They are powerful but very expensive to train and use. The extremely high inference cost, in both time and memory, is a big bottleneck for adopting a powerful transformer for solving real-world tasks at scale. Why is it hard to run inference for large transformer models? Besides the increasing size of SoTA models, there are two main factors
Neural networks are well known to be over-parameterized and can often easily fit data with near-zero training loss with decent generalization performance on test dataset. Although all these parameters are initialized at random, the optimization process can consistently lead to similarly good outcomes. And this is true even when the number of model parameters exceeds the number of training data points. Neural tangent kernel (NTK) (Jacot et al. 2018) is a kernel to explain the evolution of neural
Processing images to generate text, such as image captioning and visual question-answering, has been studied for years. Traditionally such systems rely on an object detection network as a vision encoder to capture visual features and then produce text via a text decoder. Given a large amount of existing literature, in this post, I would like to only focus on one approach for solving vision language tasks, which is to extend pre-trained generalized language models to be capable of consuming visua
Here comes the Part 3 on learning with not enough data (Previous: Part 1 and Part 2). Let’s consider two approaches for generating synthetic data for training. Augmented data. Given a set of existing training samples, we can apply a variety of augmentation, distortion and transformation to derive new data points without losing the key attributes. We have covered a bunch of augmentation methods on text and images in a previous post on contrastive learning. For the sake of post completeness, I du
This is part 2 of what to do when facing a limited amount of labeled data for supervised learning tasks. This time we will get some amount of human labeling work involved, but within a budget limit, and therefore we need to be smart when selecting which samples to label.
When facing a limited amount of labeled data for supervised learning tasks, four approaches are commonly discussed.
[Updated on 2022-03-13: add expert choice routing.] [Updated on 2022-06-10]: Greg and I wrote a shorted and upgraded version of this post, published on OpenAI Blog: “Techniques for Training Large Neural Networks”
[Updated on 2021-09-19: Highly recommend this blog post on score-based generative modeling by Yang Song (author of several key papers in the references)]. [Updated on 2022-08-27: Added classifier-free guidance, GLIDE, unCLIP and Imagen. [Updated on 2022-08-31: Added latent diffusion model. [Updated on 2024-04-13: Added progressive distillation, consistency models, and the Model Architecture section.
The goal of contrastive representation learning is to learn such an embedding space in which similar sample pairs stay close to each other while dissimilar ones are far apart. Contrastive learning can be applied to both supervised and unsupervised settings. When working with unsupervised data, contrastive learning is one of the most powerful approaches in self-supervised learning.
Large pretrained language models are trained over a sizable collection of online data. They unavoidably acquire certain toxic behavior and biases from the Internet. Pretrained language models are very powerful and have shown great success in many NLP tasks. However, to safely deploy them for practical real-world applications demands a strong safety control over the model generation process.
[Updated on 2021-02-01: Updated to version 2.0 with several work added and many typos fixed.] [Updated on 2021-05-26: Add P-tuning and Prompt Tuning in the “prompt design” section.] [Updated on 2021-09-19: Add “unlikelihood training”.]
[Updated on 2020-11-12: add an example on closed-book factual QA using OpenAI API (beta). A model that can answer any question with regard to factual knowledge can lead to many useful and practical applications, such as working as a chatbot or an AI assistant🤖. In this post, we will review several common approaches for building such an open-domain question answering system.
Although most popular and successful model architectures are designed by human experts, it doesn’t mean we have explored the entire network architecture space and settled down with the best option. We would have a better chance to find the optimal solution if we adopt a systematic and automatic way of learning high-performance model architectures.
[Updated on 2020-06-17: Add “exploration via disagreement” in the “Forward Dynamics” section. Exploitation versus exploration is a critical topic in Reinforcement Learning. We’d like the RL agent to find the best solution as fast as possible. However, in the meantime, committing to solutions too quickly without enough exploration sounds pretty bad, as it could lead to local minima or total failure. Modern RL algorithms that optimize for the best returns can achieve
[Updated on 2023-01-27: After almost three years, I did a big refactoring update of this post to incorporate a bunch of new Transformer models since 2020. The enhanced version of this post is here: The Transformer Family Version 2.0. Please refer to that post on this topic.]
[Updated on 2020-02-03: mentioning PCG in the “Task-Specific Curriculum” section. [Updated on 2020-02-04: Add a new “curriculum through distillation” section.
[Updated on 2020-01-09: add a new section on Contrastive Predictive Coding]. [Updated on 2020-04-13: add a “Momentum Contrast” section on MoCo, SimCLR and CURL.] [Updated on 2020-07-08: add a “Bisimulation” section on DeepMDP and DBC.] [Updated on 2020-09-12: add MoCo V2 and BYOL in the “Momentum Contrast” section.] [Updated on 2021-05-31: remove section on “Momentum Contrast” and add a pointer to a full post on “Contrastive Representati
Stochastic gradient descent is a universal choice for optimizing deep learning models. However, it is not the only option. With black-box optimization algorithms, you can evaluate a target function $f(x): \mathbb{R}^n \to \mathbb{R}$, even when you don’t know the precise analytic form of $f(x)$ and thus cannot compute gradients or the Hessian matrix. Examples of black-box optimization methods include Simulated Annealing, Hill Climbing and Nelder-Mead method.
In my earlier post on meta-learning, the problem is mainly defined in the context of few-shot classification. Here I would like to explore more into cases when we try to “meta-learn” Reinforcement Learning (RL) tasks by developing an agent that can solve unseen tasks fast and efficiently.
In Robotics, one of the hardest problems is how to make your model transfer to the real world. Due to the sample inefficiency of deep RL algorithms and the cost of data collection on real robots, we often need to train models in a simulator which theoretically provides an infinite amount of data. However, the reality gap between the simulator and the physical world often leads to failure when working with physical robots. The gap is triggered by an inconsistency between physical parameters (i.e
[Updated on 2019-05-27: add the section on Lottery Ticket Hypothesis.] If you are like me, entering into the field of deep learning with experience in traditional machine learning, you may often ponder over this question: Since a typical deep neural network has so many parameters and training error can easily be perfect, it should surely suffer from substantial overfitting. How could it be ever generalized to out-of-sample data points?