Since the debut of ChatGPT, there has been a notable increase in research on Large Language Models (LLMs) across a broad range of disciplines, made possible by the accessibility of this technology to a diverse user base. This fastly growing field has largely pursued two distinct paths: one aims at either scaling the model dimensions or the training dataset (or both) to enhance performance, while the other concentrates on refining smaller models (ranging from 1B to 7B parameters) with high-quality data. Despite these advances, investigations into the structural modifications of the transformer architecture itself have been relatively overlooked. Recent studies challenge the necessity of perpetually increasing model sizes by demonstrating that the deeper layers of LLMs may have minimal influence on predictive outcomes. In this work, we explore modifications to the decoder-only transformer architecture to address current challenges in the scalability and practical application of Large Language Models (LLMs).

Large Language Models (LLMs) have emerged as powerful candidates to inform clinical decision-making processes. While these models play an increasingly prominent role in shaping the digital landscape, two growing concerns emerge in healthcare applications: 1) to what extent do LLMs exhibit social bias based on patients' protected attributes (like race), and 2) how do design choices (like architecture design and prompting strategies) influence the observed biases? To answer these questions rigorously, we evaluated eight popular LLMs across three question-answering (QA) datasets using clinical vignettes (patient descriptions) standardized for bias evaluations. We employ red-teaming strategies to analyze how demographics affect LLM outputs, comparing both general-purpose and clinically-trained models. Our extensive experiments reveal various disparities (some significant) across protected groups. We also observe several counter-intuitive patterns such as larger models not being necessarily less biased and fined-tuned models on medical data not being necessarily better than the general-purpose models. Furthermore, our study demonstrates the impact of prompt design on bias patterns and shows that specific phrasing can influence bias patterns and reflection-type approaches (like Chain of Thought) can reduce biased outcomes effectively. Consistent with prior studies, we call on additional evaluations, scrutiny, and enhancement of LLMs used in clinical decision support applications.

Despite their general capabilities, LLMs still struggle on biomedical NER tasks, which are difficult due to the presence of specialized terminology and lack of training data. In this work we set out to improve LLM performance on biomedical NER in limited data settings via a new knowledge augmentation approach which incorporates definitions of relevant concepts on-the-fly. During this process, to provide a test bed for knowledge augmentation, we perform a comprehensive exploration of prompting strategies. Our experiments show that definition augmentation is useful for both open source and closed LLMs. For example, it leads to a relative improvement of 15\% (on average) in GPT-4 performance (F1) across all (six) of our test datasets. We conduct extensive ablations and analyses to demonstrate that our performance improvements stem from adding relevant definitional knowledge. We find that careful prompting strategies also improve LLM performance, allowing them to outperform fine-tuned language models in few-shot settings. To facilitate future research in this direction, we release our code at https://github.com/allenai/beacon.

Chain-of-thought (CoT) prompting can guide language models to engage in complex multi-step reasoning. The quality of provided demonstrations significantly impacts the success of downstream inference tasks. While existing automated methods prioritize accuracy and semantics in these demonstrations, we show that the underlying reasoning patterns play a more crucial role in such tasks. In this paper, we propose Pattern-Aware CoT, a prompting method that considers the diversity of demonstration patterns. By incorporating patterns such as step length and reasoning process within intermediate steps, PA-CoT effectively mitigates the issue of bias induced by demonstrations and enables better generalization to diverse scenarios. We conduct experiments on nine reasoning benchmark tasks using two open-source LLMs. The results show that our method substantially enhances reasoning performance and exhibits robustness to errors. The code will be made publicly available.

Large Vision-Language Models (LVLMs) are susceptible to object hallucinations, an issue in which their generated text contains non-existent objects, greatly limiting their reliability and practicality. Current approaches often rely on the model's token likelihoods or other internal information, instruction tuning on additional datasets, or incorporating complex external tools. We first perform empirical analysis on sentence-level LVLM hallucination, finding that CLIP similarity to the image acts as a stronger and more robust indicator of hallucination compared to token likelihoods. Motivated by this, we introduce our CLIP-Guided Decoding (CGD) approach, a straightforward but effective training-free approach to reduce object hallucination at decoding time. CGD uses CLIP to guide the model's decoding process by enhancing visual grounding of generated text with the image. Experiments demonstrate that CGD effectively mitigates object hallucination across multiple LVLM families while preserving the utility of text generation.

The web has become so interwoven with everyday life that it is easy to forget what an extraordinary accomplishment and treasure it is. In just a few decades, much of human knowledge has been collectively written up and made available to anyone with an internet connection. But all of this is coming to an end. The advent of AI threatens to destroy the complex online ecosystem that allows writers, artists, and other creators to reach human audiences. To understand why, you must understand publishing.

Large Language Models (LLMs) have experienced widespread adoption across scientific and industrial domains due to their versatility and utility for diverse tasks. Nevertheless, deploying and serving these models at scale with optimal throughput and latency remains a significant challenge, primarily because of the high computational and memory demands associated with LLMs. To tackle this limitation, we introduce Expert Router, a system designed to orchestrate multiple expert models efficiently, thereby enhancing scalability. Expert Router is a parallel inference system with a central routing gateway that distributes incoming requests using a clustering method. This approach effectively partitions incoming requests among available LLMs, maximizing overall throughput. Our extensive evaluations encompassed up to 1,000 concurrent users, providing comprehensive insights into the system's behavior from user and infrastructure perspectives. The results demonstrate Expert Router's effectiveness in handling high-load scenarios and achieving higher throughput rates, particularly under many concurrent users.

Labeling corpora constitutes a bottleneck to create models for new tasks or domains. Large language models mitigate the issue with automatic corpus labeling methods, particularly for categorical annotations. Some NLP tasks such as emotion intensity prediction, however, require text regression, but there is no work on automating annotations for continuous label assignments. Regression is considered more challenging than classification: The fact that humans perform worse when tasked to choose values from a rating scale lead to comparative annotation methods, including best-worst scaling. This raises the question if large language model-based annotation methods show similar patterns, namely that they perform worse on rating scale annotation tasks than on comparative annotation tasks. To study this, we automate emotion intensity predictions and compare direct rating scale predictions, pairwise comparisons and best-worst scaling. We find that the latter shows the highest reliability. A transformer regressor fine-tuned on these data performs nearly on par with a model trained on the original manual annotations.

Large Language Models (LLMs) are essential tools to collaborate with users on different tasks. Evaluating their performance to serve users' needs in real-world scenarios is important. While many benchmarks have been created, they mainly focus on specific predefined model abilities. Few have covered the intended utilization of LLMs by real users. To address this oversight, we propose benchmarking LLMs from a user perspective in both dataset construction and evaluation designs. We first collect 1846 real-world use cases with 15 LLMs from a user study with 712 participants from 23 countries. These self-reported cases form the User Reported Scenarios(URS) dataset with a categorization of 7 user intents. Secondly, on this authentic multi-cultural dataset, we benchmark 10 LLM services on their efficacy in satisfying user needs. Thirdly, we show that our benchmark scores align well with user-reported experience in LLM interactions across diverse intents, both of which emphasize the overlook of subjective scenarios. In conclusion, our study proposes to benchmark LLMs from a user-centric perspective, aiming to facilitate evaluations that better reflect real user needs. The benchmark dataset and code are available at https://github.com/Alice1998/URS.

This paper outlines our submission to the MEDIQA2024 Multilingual and Multimodal Medical Answer Generation (M3G) shared task. We report results for two standalone solutions under the English category of the task, the first involving two consecutive API calls to the Claude 3 Opus API and the second involving training an image-disease label joint embedding in the style of CLIP for image classification. These two solutions scored 1st and 2nd place respectively on the competition leaderboard, substantially outperforming the next best solution. Additionally, we discuss insights gained from post-competition experiments. While the performance of these two solutions have significant room for improvement due to the difficulty of the shared task and the challenging nature of medical visual question answering in general, we identify the multi-stage LLM approach and the CLIP image classification approach as promising avenues for further investigation.