The alignment of large language models is usually done by model providers to add or control behaviors that are common or universally understood across use cases and contexts. In contrast, in this article, we present an approach and architecture that empowers application developers to tune a model to their particular values, social norms, laws and other regulations, and orchestrate between potentially conflicting requirements in context. We lay out three main components of such an Alignment Studio architecture: Framers, Instructors, and Auditors that work in concert to control the behavior of a language model. We illustrate this approach with a running example of aligning a company's internal-facing enterprise chatbot to its business conduct guidelines.

Extensive efforts in the past have been directed toward the development of summarization datasets. However, a predominant number of these resources have been (semi)-automatically generated, typically through web data crawling, resulting in subpar resources for training and evaluating summarization systems, a quality compromise that is arguably due to the substantial costs associated with generating ground-truth summaries, particularly for diverse languages and specialized domains. To address this issue, we present ACLSum, a novel summarization dataset carefully crafted and evaluated by domain experts. In contrast to previous datasets, ACLSum facilitates multi-aspect summarization of scientific papers, covering challenges, approaches, and outcomes in depth. Through extensive experiments, we evaluate the quality of our resource and the performance of models based on pretrained language models and state-of-the-art large language models (LLMs). Additionally, we explore the effectiveness of extractive versus abstractive summarization within the scholarly domain on the basis of automatically discovered aspects. Our results corroborate previous findings in the general domain and indicate the general superiority of end-to-end aspect-based summarization. Our data is released at https://github.com/sobamchan/aclsum.

Writing commit messages is a tedious daily task for many software developers, and often remains neglected. Automating this task has the potential to save time while ensuring that messages are informative. A high-quality dataset and an objective benchmark are vital preconditions for solid research and evaluation towards this goal. We show that existing datasets exhibit various problems, such as the quality of the commit selection, small sample sizes, duplicates, privacy issues, and missing licenses for redistribution. This can lead to unusable models and skewed evaluations, where inferior models achieve higher evaluation scores due to biases in the data. We compile a new large-scale dataset, CommitBench, adopting best practices for dataset creation. We sample commits from diverse projects with licenses that permit redistribution and apply our filtering and dataset enhancements to improve the quality of generated commit messages. We use CommitBench to compare existing models and show that other approaches are outperformed by a Transformer model pretrained on source code. We hope to accelerate future research by publishing the source code( https://github.com/Maxscha/commitbench ).

Attribution scores indicate the importance of different input parts and can, thus, explain model behaviour. Currently, prompt-based models are gaining popularity, i.a., due to their easier adaptability in low-resource settings. However, the quality of attribution scores extracted from prompt-based models has not been investigated yet. In this work, we address this topic by analyzing attribution scores extracted from prompt-based models w.r.t. plausibility and faithfulness and comparing them with attribution scores extracted from fine-tuned models and large language models. In contrast to previous work, we introduce training size as another dimension into the analysis. We find that using the prompting paradigm (with either encoder-based or decoder-based models) yields more plausible explanations than fine-tuning the models in low-resource settings and Shapley Value Sampling consistently outperforms attention and Integrated Gradients in terms of leading to more plausible and faithful explanations.

The advancement of natural language processing (NLP) systems in healthcare hinges on language model ability to interpret the intricate information contained within clinical notes. This process often requires integrating information from various time points in a patient's medical history. However, most earlier clinical language models were pretrained with a context length limited to roughly one clinical document. In this study, We introduce ClinicalMamba, a specialized version of the Mamba language model, pretrained on a vast corpus of longitudinal clinical notes to address the unique linguistic characteristics and information processing needs of the medical domain. ClinicalMamba, with 130 million and 2.8 billion parameters, demonstrates a superior performance in modeling clinical language across extended text lengths compared to Mamba and clinical Llama. With few-shot learning, ClinicalMamba achieves notable benchmarks in speed and accuracy, outperforming existing clinical language models and general domain large models like GPT-4 in longitudinal clinical notes information extraction tasks.

Vision-Language Models (VLMs) such as GPT-4V have recently demonstrated incredible strides on diverse vision language tasks. We dig into vision-based deductive reasoning, a more sophisticated but less explored realm, and find previously unexposed blindspots in the current SOTA VLMs. Specifically, we leverage Raven's Progressive Matrices (RPMs), to assess VLMs' abilities to perform multi-hop relational and deductive reasoning relying solely on visual clues. We perform comprehensive evaluations of several popular VLMs employing standard strategies such as in-context learning, self-consistency, and Chain-of-thoughts (CoT) on three diverse datasets, including the Mensa IQ test, IntelligenceTest, and RAVEN. The results reveal that despite the impressive capabilities of LLMs in text-based reasoning, we are still far from achieving comparable proficiency in visual deductive reasoning. We found that certain standard strategies that are effective when applied to LLMs do not seamlessly translate to the challenges presented by visual reasoning tasks. Moreover, a detailed analysis reveals that VLMs struggle to solve these tasks mainly because they are unable to perceive and comprehend multiple, confounding abstract patterns in RPM examples.

Brief hospital course (BHC) summaries are common clinical documents generated by summarizing clinical notes. While large language models (LLMs) depict remarkable capabilities in automating real-world tasks, their capabilities for healthcare applications such as BHC synthesis have not been shown. To enable the adaptation of LLMs for BHC synthesis, we introduce a novel benchmark consisting of a pre-processed dataset extracted from MIMIC-IV notes, encapsulating clinical note, and brief hospital course (BHC) pairs. We assess the performance of two general-purpose LLMs and three healthcare-adapted LLMs to improve BHC synthesis from clinical notes. Using clinical notes as input for generating BHCs, we apply prompting-based (using in-context learning) and fine-tuning-based adaptation strategies to three open-source LLMs (Clinical-T5-Large, Llama2-13B, FLAN-UL2) and two proprietary LLMs (GPT-3.5, GPT-4). We quantitatively evaluate the performance of these LLMs across varying context-length inputs using conventional natural language similarity metrics. We further perform a qualitative study where five diverse clinicians blindly compare clinician-written BHCs and two LLM-generated BHCs for 30 samples across metrics of comprehensiveness, conciseness, factual correctness, and fluency. Overall, we present a new benchmark and pre-processed dataset for using LLMs in BHC synthesis from clinical notes. We observe high-quality summarization performance for both in-context proprietary and fine-tuned open-source LLMs using both quantitative metrics and a qualitative clinical reader study. We propose our work as a benchmark to motivate future works to adapt and assess the performance of LLMs in BHC synthesis.

Large Language Models (LLMs) have catalyzed transformative advances across a spectrum of natural language processing tasks through few-shot or zero-shot prompting, bypassing the need for parameter tuning. While convenient, this modus operandi aggravates ``hallucination'' concerns, particularly given the enigmatic ``black-box'' nature behind their gigantic model sizes. Such concerns are exacerbated in high-stakes applications (e.g., healthcare), where unaccountable decision errors can lead to devastating consequences. In contrast, human decision-making relies on nuanced cognitive processes, such as the ability to sense and adaptively correct misjudgments through conceptual understanding. Drawing inspiration from human cognition, we propose an innovative \textit{metacognitive} approach, dubbed \textbf{CLEAR}, to equip LLMs with capabilities for self-aware error identification and correction. Our framework facilitates the construction of concept-specific sparse subnetworks that illuminate transparent decision pathways. This provides a novel interface for model \textit{intervention} after deployment. Our intervention offers compelling advantages: (\textit{i})~at deployment or inference time, our metacognitive LLMs can self-consciously identify potential mispredictions with minimum human involvement, (\textit{ii})~the model has the capability to self-correct its errors efficiently, obviating the need for additional tuning, and (\textit{iii})~the rectification procedure is not only self-explanatory but also user-friendly, enhancing the interpretability and accessibility of the model. By integrating these metacognitive features, our approach pioneers a new path toward engendering greater trustworthiness and accountability in the deployment of LLMs.

Large language models (LLMs) are demonstrating remarkable capabilities across various tasks despite lacking a foundation in human cognition. This raises the question: can these models, beyond simply mimicking human language patterns, offer insights into the mechanisms underlying human cognition? This study explores the ability of ChatGPT to predict human performance in a language-based memory task. Building upon theories of text comprehension, we hypothesize that recognizing ambiguous sentences (e.g., "Because Bill drinks wine is never kept in the house") is facilitated by preceding them with contextually relevant information. Participants, both human and ChatGPT, were presented with pairs of sentences. The second sentence was always a garden-path sentence designed to be inherently ambiguous, while the first sentence either provided a fitting (e.g., "Bill has chronic alcoholism") or an unfitting context (e.g., "Bill likes to play golf"). We measured both human's and ChatGPT's ratings of sentence relatedness, ChatGPT's memorability ratings for the garden-path sentences, and humans' spontaneous memory for the garden-path sentences. The results revealed a striking alignment between ChatGPT's assessments and human performance. Sentences deemed more related and assessed as being more memorable by ChatGPT were indeed better remembered by humans, even though ChatGPT's internal mechanisms likely differ significantly from human cognition. This finding, which was confirmed with a robustness check employing synonyms, underscores the potential of generative AI models to predict human performance accurately. We discuss the broader implications of these findings for leveraging LLMs in the development of psychological theories and for gaining a deeper understanding of human cognition.

Retrieval-augmented generation (RAG) can enhance the generation quality of large language models (LLMs) by incorporating external token databases. However, retrievals from large databases can constitute a substantial portion of the overall generation time, particularly when retrievals are periodically performed to align the retrieved content with the latest states of generation. In this paper, we introduce PipeRAG, a novel algorithm-system co-design approach to reduce generation latency and enhance generation quality. PipeRAG integrates (1) pipeline parallelism to enable concurrent retrieval and generation processes, (2) flexible retrieval intervals to maximize the efficiency of pipeline parallelism, and (3) a performance model to automatically balance retrieval quality and latency based on the generation states and underlying hardware. Our evaluation shows that, by combining the three aforementioned methods, PipeRAG achieves up to 2.6$\times$ speedup in end-to-end generation latency while improving generation quality. These promising results showcase the effectiveness of co-designing algorithms with underlying systems, paving the way for the adoption of PipeRAG in future RAG systems.