Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context

In this report, we present the latest model of the Gemini family, Gemini 1.5 Pro, a highly compute-efficient multimodal mixture-of-experts model capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. Gemini 1.5 Pro achieves near-perfect recall on long-context retrieval tasks across modalities, improves the state-of-the-art in long-document QA, long-video QA and long-context ASR, and matches or surpasses Gemini 1.0 Ultra's state-of-the-art performance across a broad set of benchmarks. Studying the limits of Gemini 1.5 Pro's long-context ability, we find continued improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens, a generational leap over existing models such as Claude 2.1 (200k) and GPT-4 Turbo (128k). Finally, we highlight surprising new capabilities of large language models at the frontier; when given a grammar manual for Kalamang, a language with fewer than 200 speakers worldwide, the model learns to translate English to Kalamang at a similar level to a person who learned from the same content.

Gemini: A Family of Highly Capable Multimodal Models

This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra model advances the state of the art in 30 of 32 of these benchmarks - notably being the first model to achieve human-expert performance on the well-studied exam benchmark MMLU, and improving the state of the art in every one of the 20 multimodal benchmarks we examined. We believe that the new capabilities of Gemini models in cross-modal reasoning and language understanding will enable a wide variety of use cases and we discuss our approach toward deploying them responsibly to users.

Exploring Length Generalization in Large Language Models

The ability to extrapolate from short problem instances to longer ones is an important form of out-of-distribution generalization in reasoning tasks, and is crucial when learning from datasets where longer problem instances are rare. These include theorem proving, solving quantitative mathematics problems, and reading/summarizing novels. In this paper, we run careful empirical studies exploring the length generalization capabilities of transformer-based language models. We first establish that naively finetuning transformers on length generalization tasks shows significant generalization deficiencies independent of model scale. We then show that combining pretrained large language models' in-context learning abilities with scratchpad prompting (asking the model to output solution steps before producing an answer) results in a dramatic improvement in length generalization. We run careful failure analyses on each of the learning modalities and identify common sources of mistakes that highlight opportunities in equipping language models with the ability to generalize to longer problems.

Solving Quantitative Reasoning Problems with Language Models

Language models have achieved remarkable performance on a wide range of tasks that require natural language understanding. Nevertheless, state-of-the-art models have generally struggled with tasks that require quantitative reasoning, such as solving mathematics, science, and engineering problems at the college level. To help close this gap, we introduce Minerva, a large language model pretrained on general natural language data and further trained on technical content. The model achieves state-of-the-art performance on technical benchmarks without the use of external tools. We also evaluate our model on over two hundred undergraduate-level problems in physics, biology, chemistry, economics, and other sciences that require quantitative reasoning, and find that the model can correctly answer nearly a third of them.

Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models

Language models demonstrate both quantitative improvement and new qualitative capabilities with increasing scale. Despite their potentially transformative impact, these new capabilities are as yet poorly characterized. In order to inform future research, prepare for disruptive new model capabilities, and ameliorate socially harmful effects, it is vital that we understand the present and near-future capabilities and limitations of language models. To address this challenge, we introduce the Beyond the Imitation Game benchmark (BIG-bench). BIG-bench currently consists of 204 tasks, contributed by 442 authors across 132 institutions. Task topics are diverse, drawing problems from linguistics, childhood development, math, common-sense reasoning, biology, physics, social bias, software development, and beyond. BIG-bench focuses on tasks that are believed to be beyond the capabilities of current language models. We evaluate the behavior of OpenAI's GPT models, Google-internal dense transformer architectures, and Switch-style sparse transformers on BIG-bench, across model sizes spanning millions to hundreds of billions of parameters. In addition, a team of human expert raters performed all tasks in order to provide a strong baseline. Findings include: model performance and calibration both improve with scale, but are poor in absolute terms (and when compared with rater performance); performance is remarkably similar across model classes, though with benefits from sparsity; tasks that improve gradually and predictably commonly involve a large knowledge or memorization component, whereas tasks that exhibit "breakthrough" behavior at a critical scale often involve multiple steps or components, or brittle metrics; social bias typically increases with scale in settings with ambiguous context, but this can be improved with prompting.

The Evolution of Out-of-Distribution Robustness Throughout Fine-Tuning

Although machine learning models typically experience a drop in performance on out-of-distribution data, accuracies on in- versus out-of-distribution data are widely observed to follow a single linear trend when evaluated across a testbed of models. Models that are more accurate on the out-of-distribution data relative to this baseline exhibit "effective robustness" and are exceedingly rare. Identifying such models, and understanding their properties, is key to improving out-of-distribution performance. We conduct a thorough empirical investigation of effective robustness during fine-tuning and surprisingly find that models pre-trained on larger datasets exhibit effective robustness during training that vanishes at convergence. We study how properties of the data influence effective robustness, and we show that it increases with the larger size, more diversity, and higher example difficulty of the dataset. We also find that models that display effective robustness are able to correctly classify 10% of the examples that no other current testbed model gets correct. Finally, we discuss several strategies for scaling effective robustness to the high-accuracy regime to improve the out-of-distribution accuracy of state-of-the-art models.

Scaffolding Simulations with Deep Learning for High-dimensional Deconvolution

A common setting for scientific inference is the ability to sample from a high-fidelity forward model (simulation) without having an explicit probability density of the data. We propose a simulation-based maximum likelihood deconvolution approach in this setting called OmniFold. Deep learning enables this approach to be naturally unbinned and (variable-, and) high-dimensional. In contrast to model parameter estimation, the goal of deconvolution is to remove detector distortions in order to enable a variety of down-stream inference tasks. Our approach is the deep learning generalization of the common Richardson-Lucy approach that is also called Iterative Bayesian Unfolding in particle physics. We show how OmniFold can not only remove detector distortions, but it can also account for noise processes and acceptance effects.

Understanding the Failure Modes of Out-of-Distribution Generalization

Empirical studies suggest that machine learning models often rely on features, such as the background, that may be spuriously correlated with the label only during training time, resulting in poor accuracy during test-time. In this work, we identify the fundamental factors that give rise to this behavior, by explaining why models fail this way {\em even} in easy-to-learn tasks where one would expect these models to succeed. In particular, through a theoretical study of gradient-descent-trained linear classifiers on some easy-to-learn tasks, we uncover two complementary failure modes. These modes arise from how spurious correlations induce two kinds of skews in the data: one geometric in nature, and another, statistical in nature. Finally, we construct natural modifications of image classification datasets to understand when these failure modes can arise in practice. We also design experiments to isolate the two failure modes when training modern neural networks on these datasets.

Parameter Estimation using Neural Networks in the Presence of Detector Effects

Histogram-based template fits are the main technique used for estimating parameters of high energy physics Monte Carlo generators. Parameterized neural network reweighting can be used to extend this fitting procedure to many dimensions and does not require binning. If the fit is to be performed using reconstructed data, then expensive detector simulations must be used for training the neural networks. We introduce a new two-level fitting approach that only requires one dataset with detector simulation and then a set of additional generation-level datasets without detector effects included. This Simulation-level fit based on Reweighting Generator-level events with Neural networks (SRGN) is demonstrated using simulated datasets for a variety of examples including a simple Gaussian random variable, parton shower tuning, and the top quark mass extraction.

Asymptotics of Wide Convolutional Neural Networks

Wide neural networks have proven to be a rich class of architectures for both theory and practice. Motivated by the observation that finite width convolutional networks appear to outperform infinite width networks, we study scaling laws for wide CNNs and networks with skip connections. Following the approach of (Dyer & Gur-Ari, 2019), we present a simple diagrammatic recipe to derive the asymptotic width dependence for many quantities of interest. These scaling relationships provide a solvable description for the training dynamics of wide convolutional networks. We test these relations across a broad range of architectures. In particular, we find that the difference in performance between finite and infinite width models vanishes at a definite rate with respect to model width. Nonetheless, this relation is consistent with finite width models generalizing either better or worse than their infinite width counterparts, and we provide examples where the relative performance depends on the optimization details.