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.

Modeling subjectivity (by Mimicking Annotator Annotation) in toxic comment identification across diverse communities

The prevalence and impact of toxic discussions online have made content moderation crucial.Automated systems can play a vital role in identifying toxicity, and reducing the reliance on human moderation.Nevertheless, identifying toxic comments for diverse communities continues to present challenges that are addressed in this paper.The two-part goal of this study is to(1)identify intuitive variances from annotator disagreement using quantitative analysis and (2)model the subjectivity of these viewpoints.To achieve our goal, we published a new dataset\footnote{\url{https://github.com/XXX}} with expert annotators' annotations and used two other public datasets to identify the subjectivity of toxicity.Then leveraging the Large Language Model(LLM),we evaluate the model's ability to mimic diverse viewpoints on toxicity by varying size of the training data and utilizing same set of annotators as the test set used during model training and a separate set of annotators as the test set.We conclude that subjectivity is evident across all annotator groups, demonstrating the shortcomings of majority-rule voting. Moving forward, subjective annotations should serve as ground truth labels for training models for domains like toxicity in diverse communities.

Using Foundation Models to Detect Policy Violations with Minimal Supervision

Foundation models, i.e. large neural networks pre-trained on large text corpora, have revolutionized NLP. They can be instructed directly (e.g. (arXiv:2005.14165)) - this is called hard prompting - and they can be tuned using very little data (e.g. (arXiv:2104.08691)) - this technique is called soft prompting. We seek to leverage their capabilities to detect policy violations. Our contributions are: We identify a hard prompt that adapts chain-of-thought prompting to policy violation tasks. This prompt produces policy violation classifications, along with extractive explanations that justify the classification. We compose the hard-prompts with soft prompt tuning to produce a classifier that attains high accuracy with very little supervision; the same classifier also produces explanations. Though the supervision only acts on the classifications, we find that the modified explanations remain consistent with the (tuned) model's response. Along the way, we identify several unintuitive aspects of foundation models. For instance, adding an example from a specific class can actually reduce predictions of that class, and separately, the effects of tokenization on scoring etc. Based on our technical results, we identify a simple workflow for product teams to quickly develop effective policy violation detectors.

Understanding Finetuning for Factual Knowledge Extraction from Language Models

Language models (LMs) pretrained on large corpora of text from the web have been observed to contain large amounts of various types of knowledge about the world. This observation has led to a new and exciting paradigm in knowledge graph construction where, instead of manual curation or text mining, one extracts knowledge from the parameters of an LM. Recently, it has been shown that finetuning LMs on a set of factual knowledge makes them produce better answers to queries from a different set, thus making finetuned LMs a good candidate for knowledge extraction and, consequently, knowledge graph construction. In this paper, we analyze finetuned LMs for factual knowledge extraction. We show that along with its previously known positive effects, finetuning also leads to a (potentially harmful) phenomenon which we call Frequency Shock, where at the test time the model over-predicts rare entities that appear in the training set and under-predicts common entities that do not appear in the training set enough times. We show that Frequency Shock leads to a degradation in the predictions of the model and beyond a point, the harm from Frequency Shock can even outweigh the positive effects of finetuning, making finetuning harmful overall. We then consider two solutions to remedy the identified negative effect: 1- model mixing and 2- mixture finetuning with the LM's pre-training task. The two solutions combined lead to significant improvements compared to vanilla finetuning.

Robust Handling of Polysemy via Sparse Representations

Words are polysemous and multi-faceted, with many shades of meanings. We suggest that sparse distributed representations are more suitable than other, commonly used, (dense) representations to express these multiple facets, and present Category Builder, a working system that, as we show, makes use of sparse representations to support multi-faceted lexical representations. We argue that the set expansion task is well suited to study these meaning distinctions since a word may belong to multiple sets with a different reason for membership in each. We therefore exhibit the performance of Category Builder on this task, while showing that our representation captures at the same time analogy problems such as "the Ganga of Egypt" or "the Voldemort of Tolkien". Category Builder is shown to be a more expressive lexical representation and to outperform dense representations such as Word2Vec in some analogy classes despite being shown only two of the three input terms.