Recent advances in large multimodal models (LMMs) suggest that higher image resolution enhances the fine-grained understanding of image details, crucial for tasks such as visual commonsense reasoning and analyzing biomedical images. However, increasing input resolution poses two main challenges: 1) It extends the context length required by the language model, leading to inefficiencies and hitting the model's context limit; 2) It increases the complexity of visual features, necessitating more training data or more complex architecture. We introduce Dragonfly, a new LMM architecture that enhances fine-grained visual understanding and reasoning about image regions to address these challenges. Dragonfly employs two key strategies: multi-resolution visual encoding and zoom-in patch selection. These strategies allow the model to process high-resolution images efficiently while maintaining reasonable context length. Our experiments on eight popular benchmarks demonstrate that Dragonfly achieves competitive or better performance compared to other architectures, highlighting the effectiveness of our design. Additionally, we finetuned Dragonfly on biomedical instructions, achieving state-of-the-art results on multiple biomedical tasks requiring fine-grained visual understanding, including 92.3% accuracy on the Path-VQA dataset (compared to 83.3% for Med-Gemini) and the highest reported results on biomedical image captioning. To support model training, we curated a visual instruction-tuning dataset with 5.5 million image-instruction samples in the general domain and 1.4 million samples in the biomedical domain. We also conducted ablation studies to characterize the impact of various architectural designs and image resolutions, providing insights for future research on visual instruction alignment.

The issue of popularity bias -- where popular items are disproportionately recommended, overshadowing less popular but potentially relevant items -- remains a significant challenge in recommender systems. Recent advancements have seen the integration of general-purpose Large Language Models (LLMs) into the architecture of such systems. This integration raises concerns that it might exacerbate popularity bias, given that the LLM's training data is likely dominated by popular items. However, it simultaneously presents a novel opportunity to address the bias via prompt tuning. Our study explores this dichotomy, examining whether LLMs contribute to or can alleviate popularity bias in recommender systems. We introduce a principled way to measure popularity bias by discussing existing metrics and proposing a novel metric that fulfills a series of desiderata. Based on our new metric, we compare a simple LLM-based recommender to traditional recommender systems on a movie recommendation task. We find that the LLM recommender exhibits less popularity bias, even without any explicit mitigation.

Input-output safeguards are used to detect anomalies in the traces produced by Large Language Models (LLMs) systems. These detectors are at the core of diverse safety-critical applications such as real-time monitoring, offline evaluation of traces, and content moderation. However, there is no widely recognized methodology to evaluate them. To fill this gap, we introduce the Benchmarks for the Evaluation of LLM Safeguards (BELLS), a structured collection of tests, organized into three categories: (1) established failure tests, based on already-existing benchmarks for well-defined failure modes, aiming to compare the performance of current input-output safeguards; (2) emerging failure tests, to measure generalization to never-seen-before failure modes and encourage the development of more general safeguards; (3) next-gen architecture tests, for more complex scaffolding (such as LLM-agents and multi-agent systems), aiming to foster the development of safeguards that could adapt to future applications for which no safeguard currently exists. Furthermore, we implement and share the first next-gen architecture test, using the MACHIAVELLI environment, along with an interactive visualization of the dataset.

Deductive reasoning plays a pivotal role in the formulation of sound and cohesive arguments. It allows individuals to draw conclusions that logically follow, given the truth value of the information provided. Recent progress in the domain of large language models (LLMs) has showcased their capability in executing deductive reasoning tasks. Nonetheless, a significant portion of research primarily assesses the accuracy of LLMs in solving such tasks, often overlooking a deeper analysis of their reasoning behavior. In this study, we draw upon principles from cognitive psychology to examine inferential strategies employed by LLMs, through a detailed evaluation of their responses to propositional logic problems. Our findings indicate that LLMs display reasoning patterns akin to those observed in humans, including strategies like $\textit{supposition following}$ or $\textit{chain construction}$. Moreover, our research demonstrates that the architecture and scale of the model significantly affect its preferred method of reasoning, with more advanced models tending to adopt strategies more frequently than less sophisticated ones. Importantly, we assert that a model's accuracy, that is the correctness of its final conclusion, does not necessarily reflect the validity of its reasoning process. This distinction underscores the necessity for more nuanced evaluation procedures in the field.

Fact-checking is a crucial natural language processing (NLP) task that verifies the truthfulness of claims by considering reliable evidence. Traditional methods are often limited by labour-intensive data curation and rule-based approaches. In this paper, we present FactGenius, a novel method that enhances fact-checking by combining zero-shot prompting of large language models (LLMs) with fuzzy text matching on knowledge graphs (KGs). Leveraging DBpedia, a structured linked data dataset derived from Wikipedia, FactGenius refines LLM-generated connections using similarity measures to ensure accuracy. The evaluation of FactGenius on the FactKG, a benchmark dataset for fact verification, demonstrates that it significantly outperforms existing baselines, particularly when fine-tuning RoBERTa as a classifier. The two-stage approach of filtering and validating connections proves crucial, achieving superior performance across various reasoning types and establishing FactGenius as a promising tool for robust fact-checking. The code and materials are available at https://github.com/SushantGautam/FactGenius.

Predicting the locations an individual will visit in the future is crucial for solving many societal issues like disease diffusion and reduction of pollution among many others. The models designed to tackle next-location prediction, however, require a significant amount of individual-level information to be trained effectively. Such data may be scarce or even unavailable in some geographic regions or peculiar scenarios (e.g., cold-start in recommendation systems). Moreover, the design of a next-location predictor able to generalize or geographically transfer knowledge is still an open research challenge. Recent advances in natural language processing have led to a rapid diffusion of Large Language Models (LLMs) which have shown good generalization and reasoning capabilities. These insights, coupled with the recent findings that LLMs are rich in geographical knowledge, allowed us to believe that these models can act as zero-shot next-location predictors. This paper evaluates the capabilities of many popular LLMs in this role, specifically Llama, GPT-3.5 and Mistral 7B. After designing a proper prompt, we tested the models on three real-world mobility datasets. The results show that LLMs can obtain accuracies up to 32.4%, a significant relative improvement of over 600% when compared to sophisticated DL models specifically designed for human mobility. Moreover, we show that other LLMs are unable to perform the task properly. To prevent positively biased results, we also propose a framework inspired by other studies to test data contamination. Finally, we explored the possibility of using LLMs as text-based explainers for next-location prediction showing that can effectively provide an explanation for their decision. Notably, 7B models provide more generic, but still reliable, explanations compared to larger counterparts. Code: github.com/ssai-trento/LLM-zero-shot-NL

Large language models (LLMs) often incorporate multiple text chunks in their inputs to provide the necessary contexts. To speed up the prefill of the long LLM inputs, one can pre-compute the KV cache of a text and re-use the KV cache when the context is reused as the prefix of another LLM input. However, the reused text chunks are not always the input prefix, and when they are not, their precomputed KV caches cannot be directly used since they ignore the text's cross-attention with the preceding text in the LLM input. Thus, the benefits of reusing KV caches remain largely unrealized. This paper tackles just one question: when an LLM input contains multiple text chunks, how to quickly combine their precomputed KV caches in order to achieve the same generation quality as the expensive full prefill (i.e., without reusing KV cache)? We present CacheBlend, a scheme that reuses the pre-computed KV caches, regardless prefix or not, and selectively recomputes the KV values of a small subset of tokens to partially update each reused KV cache. In the meantime,the small extra delay for recomputing some tokens can be pipelined with the retrieval of KV caches within the same job,allowing CacheBlend to store KV caches in slower devices with more storage capacity while retrieving them without increasing the inference delay. By comparing CacheBlend with the state-of-the-art KV cache reusing schemes on three open-source LLMs of various sizes and four popular benchmark datasets of different tasks, we show that CacheBlend reduces time-to-first-token (TTFT) by 2.2-3.3X and increases the inference throughput by 2.8-5X, compared with full KV recompute, without compromising generation quality or incurring more storage cost.

Decoding language from brain dynamics is an important open direction in the realm of brain-computer interface (BCI), especially considering the rapid growth of large language models. Compared to invasive-based signals which require electrode implantation surgery, non-invasive neural signals (e.g. EEG, MEG) have attracted increasing attention considering their safety and generality. However, the exploration is not adequate in three aspects: 1) previous methods mainly focus on EEG but none of the previous works address this problem on MEG with better signal quality; 2) prior works have predominantly used $``teacher-forcing"$ during generative decoding, which is impractical; 3) prior works are mostly $``BART-based"$ not fully auto-regressive, which performs better in other sequence tasks. In this paper, we explore the brain-to-text translation of MEG signals in a speech-decoding formation. Here we are the first to investigate a cross-attention-based ``whisper" model for generating text directly from MEG signals without teacher forcing. Our model achieves impressive BLEU-1 scores of 60.30 and 52.89 without pretraining $\&$ teacher-forcing on two major datasets ($\textit{GWilliams}$ and $\textit{Schoffelen}$). This paper conducts a comprehensive review to understand how speech decoding formation performs on the neural decoding tasks, including pretraining initialization, training $\&$ evaluation set splitting, augmentation, and scaling law. Code is available at https://github.com/NeuSpeech/NeuSpeech1$.

Code Large Language Models (Code LLMs) have excelled at tasks like code completion but often miss deeper semantics such as execution effects and dynamic states. This paper aims to bridge the gap between Code LLMs' reliance on static text data and the need for thorough semantic understanding for complex tasks like debugging and program repair. We introduce a novel strategy to train Code LLMs with comprehensive semantics, encompassing high-level functional descriptions, local execution effects of individual statements, and overall input/output behavior, thereby linking static code text with dynamic execution states. We begin by collecting PyX, a clean code corpus of fully executable samples with functional descriptions and execution tracing. We propose training Code LLMs to write code and represent and reason about execution behaviors using natural language, mimicking human verbal debugging. This approach led to the development of SemCoder, a Code LLM with only 6.7B parameters, which shows competitive performance with GPT-3.5-turbo on code generation and execution reasoning tasks. SemCoder achieves 81.1% on HumanEval (GPT-3.5-turbo: 76.8%) and 54.5% on CRUXEval-I (GPT-3.5-turbo: 50.3%). We also study the effectiveness of SemCoder's monologue-style execution reasoning compared to concrete scratchpad reasoning, showing that our approach integrates semantics from multiple dimensions more smoothly. Finally, we demonstrate the potential of applying learned semantics to improve Code LLMs' debugging and self-refining capabilities.

Designing effective reward functions is crucial to training reinforcement learning (RL) algorithms. However, this design is non-trivial, even for domain experts, due to the subjective nature of certain tasks that are hard to quantify explicitly. In recent works, large language models (LLMs) have been used for reward generation from natural language task descriptions, leveraging their extensive instruction tuning and commonsense understanding of human behavior. In this work, we hypothesize that LLMs, guided by human feedback, can be used to formulate human-aligned reward functions. Specifically, we study this in the challenging setting of autonomous driving (AD), wherein notions of "good" driving are tacit and hard to quantify. To this end, we introduce REvolve, an evolutionary framework that uses LLMs for reward design in AD. REvolve creates and refines reward functions by utilizing human feedback to guide the evolution process, effectively translating implicit human knowledge into explicit reward functions for training (deep) RL agents. We demonstrate that agents trained on REvolve-designed rewards align closely with human driving standards, thereby outperforming other state-of-the-art baselines.