Large language models (LLMs) are increasingly trained from AI constitutions and model specifications that establish behavioral guidelines and ethical principles. However, these specifications face critical challenges, including internal conflicts between principles and insufficient coverage of nuanced scenarios. We present a systematic methodology for stress-testing model character specifications, automatically identifying numerous cases of principle contradictions and interpretive ambiguities in current model specs. We stress test current model specs by generating scenarios that force explicit tradeoffs between competing value-based principles. Using a comprehensive taxonomy we generate diverse value tradeoff scenarios where models must choose between pairs of legitimate principles that cannot be simultaneously satisfied. We evaluate responses from twelve frontier LLMs across major providers (Anthropic, OpenAI, Google, xAI) and measure behavioral disagreement through value classification scores. Among these scenarios, we identify over 70,000 cases exhibiting significant behavioral divergence. Empirically, we show this high divergence in model behavior strongly predicts underlying problems in model specifications. Through qualitative analysis, we provide numerous example issues in current model specs such as direct contradiction and interpretive ambiguities of several principles. Additionally, our generated dataset also reveals both clear misalignment cases and false-positive refusals across all of the frontier models we study. Lastly, we also provide value prioritization patterns and differences of these models.

We investigate how agents built on pretrained large language models can learn target classification functions from labeled examples without parameter updates. While conventional approaches like fine-tuning are often costly, inflexible, and opaque, we propose a memory-augmented framework that leverages both labeled data and LLM-generated critiques. Our framework uses episodic memory to store instance-level critiques-capturing specific past experiences-and semantic memory to distill these into reusable, task-level guidance. Across a diverse set of tasks, incorporating critiques yields up to a 24.8 percent accuracy improvement over retrieval-based (RAG-style) baselines that rely only on labels. Through extensive empirical evaluation, we uncover distinct behavioral differences between OpenAI and opensource models, particularly in how they handle fact-oriented versus preference-based data. To interpret how models respond to different representations of supervision encoded in memory, we introduce a novel metric, suggestibility. This helps explain observed behaviors and illuminates how model characteristics and memory strategies jointly shape learning dynamics. Our findings highlight the promise of memory-driven, reflective learning for building more adaptive and interpretable LLM agents.

This paper explores the spatial reasoning capability of large language models (LLMs) over textual input through a suite of five tasks aimed at probing their spatial understanding and computational abilities. The models were tested on both fundamental spatial reasoning and multi-step problem-solving within structured grid-based environments using tasks such as quadrant identification, geometric transformations, distance evaluation, word searches, and tile sliding. Each task was scaled in complexity through increasing grid dimensions, requiring models to extend beyond simple pattern recognition into abstract spatial reasoning. Our results reveal that while LLMs demonstrate moderate success in all tasks with small complexity and size, performance drops off rapidly as scale increases, with an average loss in accuracy of 42.7%, and reaching as high as 84%. Every test that began with over 50% accuracy showed a loss of at least 48%, illustrating the consistent nature of the deterioration. Furthermore, their struggles with scaling complexity hint at a lack of robust spatial representations in their underlying architectures. This paper underscores the gap between linguistic and spatial reasoning in LLMs, offering insights into their current limitations, and laying the groundwork for future integrative benchmarks at the intersection of language and geometry.

AI-powered search systems are emerging as new information gatekeepers, fundamentally transforming how users access news and information. Despite their growing influence, the citation patterns of these systems remain poorly understood. We address this gap by analyzing data from the AI Search Arena, a head-to-head evaluation platform for AI search systems. The dataset comprises over 24,000 conversations and 65,000 responses from models across three major providers: OpenAI, Perplexity, and Google. Among the over 366,000 citations embedded in these responses, 9% reference news sources. We find that while models from different providers cite distinct news sources, they exhibit shared patterns in citation behavior. News citations concentrate heavily among a small number of outlets and display a pronounced liberal bias, though low-credibility sources are rarely cited. User preference analysis reveals that neither the political leaning nor the quality of cited news sources significantly influences user satisfaction. These findings reveal significant challenges in current AI search systems and have important implications for their design and governance.

Large language models (LLMs) are revolutionizing healthcare by improving diagnosis, patient care, and decision support through interactive communication. More recently, they have been applied to analyzing physiological time-series like wearable data for health insight extraction. Existing methods embed raw numerical sequences directly into prompts, which exceeds token limits and increases computational costs. Additionally, some studies integrated features extracted from time-series in textual prompts or applied multimodal approaches. However, these methods often produce generic and unreliable outputs due to LLMs' limited analytical rigor and inefficiency in interpreting continuous waveforms. In this paper, we develop an LLM-powered agent for physiological time-series analysis aimed to bridge the gap in integrating LLMs with well-established analytical tools. Built on the OpenCHA, an open-source LLM-powered framework, our agent features an orchestrator that integrates user interaction, data sources, and analytical tools to generate accurate health insights. To evaluate its effectiveness, we implement a case study on heart rate (HR) estimation from Photoplethysmogram (PPG) signals using a dataset of PPG and Electrocardiogram (ECG) recordings in a remote health monitoring study. The agent's performance is benchmarked against OpenAI GPT-4o-mini and GPT-4o, with ECG serving as the gold standard for HR estimation. Results demonstrate that our agent significantly outperforms benchmark models by achieving lower error rates and more reliable HR estimations. The agent implementation is publicly available on GitHub.

Recent advances in large language models (LLMs) have led to the development of various evaluation benchmarks. These benchmarks typically rely on a single instruction template for evaluating all LLMs on a specific task. In this paper, we comprehensively analyze the brittleness of results obtained via single-prompt evaluations across 6.5M instances, involving 20 different LLMs and 39 tasks from 3 benchmarks. To improve robustness of the analysis, we propose to evaluate LLMs with a set of diverse prompts instead. We discuss tailored evaluation metrics for specific use cases (e.g., LLM developers vs. developers interested in a specific downstream task), ensuring a more reliable and meaningful assessment of LLM capabilities. We then implement these criteria and conduct evaluations of multiple models, providing insights into the true strengths and limitations of current LLMs.

Prior work shows that program-aided reasoning, in which large language models (LLMs) are combined with programs written in programming languages such as Python, can significantly improve accuracy on various reasoning tasks. However, while accuracy is essential, it is also important for such reasoners to "know what they know", which can be quantified through the calibration of the model. In this paper, we compare the calibration of Program Aided Language Models (PAL) and text-based Chain-of-thought (COT) prompting techniques over 5 datasets and 2 model types: LLaMA models and OpenAI models. Our results indicate that PAL leads to improved calibration in 75% of the instances. Our analysis uncovers that prompting styles that produce lesser diversity in generations also have more calibrated results, and thus we also experiment with inducing lower generation diversity using temperature scaling and find that for certain temperatures, PAL is not only more accurate but is also more calibrated than COT. Overall, we demonstrate that, in the majority of cases, program-aided reasoners better know what they know than text-based counterparts.

This study presents an innovative approach to the application of large language models (LLMs) in clinical decision-making, focusing on OpenAI's ChatGPT. Our approach introduces the use of contextual prompts-strategically designed to include task description, feature description, and crucially, integration of domain knowledge-for high-quality binary classification tasks even in data-scarce scenarios. The novelty of our work lies in the utilization of domain knowledge, obtained from high-performing interpretable ML models, and its seamless incorporation into prompt design. By viewing these ML models as medical experts, we extract key insights on feature importance to aid in decision-making processes. This interplay of domain knowledge and AI holds significant promise in creating a more insightful diagnostic tool. Additionally, our research explores the dynamics of zero-shot and few-shot prompt learning based on LLMs. By comparing the performance of OpenAI's ChatGPT with traditional supervised ML models in different data conditions, we aim to provide insights into the effectiveness of prompt engineering strategies under varied data availability. In essence, this paper bridges the gap between AI and healthcare, proposing a novel methodology for LLMs application in clinical decision support systems. It highlights the transformative potential of effective prompt design, domain knowledge integration, and flexible learning approaches in enhancing automated decision-making.

OpenAI is a research and deployment company. They are the creators of the models powering experiences like ChatGPT and Bing Image Creator.

ChatGPT after its launch on November 30, 2022, took the AI world by storm. Since its launch, ChatGPT was awarded "the best AI chatbot ever released" by the New York Times. This AI chatbot which provides human-like text answers to users has broken many unexpected records since its launch. The AI bot hit its first 1 million users in only 5 days and reached the 100 million monthly active users mark in just 2 months after the launch. It also scared some big tech industries in the world to change their strategies like Google declaring code red after ChatGPT's popularity and introducing its own AI tool Bard AI.