Fine-tuning large language models (LLMs) under resource constraints is a significant challenge in deep learning. Low-Rank Adaptation (LoRA), pruning, and quantization are all effective methods for improving resource efficiency. However, combining them directly often results in suboptimal performance, especially with uniform quantization across all model layers. This is due to the complex, uneven interlayer relationships introduced by pruning, necessitating more refined quantization strategies. To address this, we propose AutoMixQ, an end-to-end optimization framework that selects optimal quantization configurations for each LLM layer. AutoMixQ leverages lightweight performance models to guide the selection process, significantly reducing time and computational resources compared to exhaustive search methods. By incorporating Pareto optimality, AutoMixQ balances memory usage and performance, approaching the upper bounds of model capability under strict resource constraints. Our experiments on widely used benchmarks show that AutoMixQ reduces memory consumption while achieving superior performance. For example, at a 30\% pruning rate in LLaMA-7B, AutoMixQ achieved 66.21\% on BoolQ compared to 62.45\% for LoRA and 58.96\% for LoftQ, while reducing memory consumption by 35.5\% compared to LoRA and 27.5\% compared to LoftQ.

Use of large language models such as ChatGPT (GPT-4) for mental health support has grown rapidly, emerging as a promising route to assess and help people with mood disorders, like depression. However, we have a limited understanding of GPT-4's schema of mental disorders, that is, how it internally associates and interprets symptoms. In this work, we leveraged contemporary measurement theory to decode how GPT-4 interrelates depressive symptoms to inform both clinical utility and theoretical understanding. We found GPT-4's assessment of depression: (a) had high overall convergent validity (r = .71 with self-report on 955 samples, and r = .81 with experts judgments on 209 samples); (b) had moderately high internal consistency (symptom inter-correlates r = .23 to .78 ) that largely aligned with literature and self-report; except that GPT-4 (c) underemphasized suicidality's -- and overemphasized psychomotor's -- relationship with other symptoms, and (d) had symptom inference patterns that suggest nuanced hypotheses (e.g. sleep and fatigue are influenced by most other symptoms while feelings of worthlessness/guilt is mostly influenced by depressed mood).

Large Language Models (LLMs) have demonstrated impressive capabilities in generating coherent text but often struggle with grounding language and strategic dialogue. To address this gap, we focus on journalistic interviews, a domain rich in grounding communication and abundant in data. We curate a dataset of 40,000 two-person informational interviews from NPR and CNN, and reveal that LLMs are significantly less likely than human interviewers to use acknowledgements and to pivot to higher-level questions. Realizing that a fundamental deficit exists in multi-turn planning and strategic thinking, we develop a realistic simulated environment, incorporating source personas and persuasive elements, in order to facilitate the development of agents with longer-horizon rewards. Our experiments show that while source LLMs mimic human behavior in information sharing, interviewer LLMs struggle with recognizing when questions are answered and engaging persuasively, leading to suboptimal information extraction across model size and capability. These findings underscore the need for enhancing LLMs' strategic dialogue capabilities.

The advent of Large Language Models (LLMs) has enabled the development of LLM agents capable of autonomously achieving under-specified goals and continuously evolving through post-deployment improvement, sometimes without requiring code or model updates. Conventional approaches, such as pre-defined test cases and code/model redevelopment pipelines, are inadequate for addressing the unique challenges of LLM agent development, particularly in terms of quality and risk control. This paper introduces an evaluation-driven design approach, inspired by test-driven development, to address these challenges. Through a multivocal literature review (MLR), we synthesize existing LLM evaluation methods and propose a novel process model and reference architecture specifically designed for LLM agents. The proposed approach integrates online and offline evaluations to support adaptive runtime adjustments and systematic offline redevelopment, improving runtime pipelines, artifacts, system architecture, and LLMs by continuously incorporating evaluation results, including fine-grained feedback from human and AI evaluators.

Large language model (LLM) evaluations often assume there is a single correct response -- a gold label -- for each item in the evaluation corpus. However, some tasks can be ambiguous -- i.e., they provide insufficient information to identify a unique interpretation -- or vague -- i.e., they do not clearly indicate where to draw the line when making a determination. Both ambiguity and vagueness can cause task indeterminacy -- the condition where some items in the evaluation corpus have more than one correct response. In this paper, we develop a framework for evaluating LLMs under task indeterminacy. Our framework disentangles the relationships between task specification, human ratings, and LLM responses in the LLM evaluation pipeline. Using our framework, we conduct a synthetic experiment showing that evaluations that use the "gold label" assumption underestimate the true performance. We also provide a method for estimating an error-adjusted performance interval given partial knowledge about indeterminate items in the evaluation corpus. We conclude by outlining implications of our work for the research community.

Evaluating the output of Large Language Models (LLMs) is one of the most critical aspects of building a performant compound AI system. Since the output from LLMs propagate to downstream steps, identifying LLM errors is crucial to system performance. A common task for LLMs in AI systems is tool use. While there are several benchmark environments for evaluating LLMs on this task, they typically only give a success rate without any explanation of the failure cases. To solve this problem, we introduce SpecTool, a new benchmark to identify error patterns in LLM output on tool-use tasks. Our benchmark data set comprises of queries from diverse environments that can be used to test for the presence of seven newly characterized error patterns. Using SPECTOOL , we show that even the most prominent LLMs exhibit these error patterns in their outputs. Researchers can use the analysis and insights from SPECTOOL to guide their error mitigation strategies.

Large Language Models (LLMs) and Vision Language Models (VLMs) possess extensive knowledge and exhibit promising reasoning abilities; however, they still struggle to perform well in complex, dynamic environments. Real-world tasks require handling intricate interactions, advanced spatial reasoning, long-term planning, and continuous exploration of new strategies-areas in which we lack effective methodologies for comprehensively evaluating these capabilities. To address this gap, we introduce BALROG, a novel benchmark designed to assess the agentic capabilities of LLMs and VLMs through a diverse set of challenging games. Our benchmark incorporates a range of existing reinforcement learning environments with varying levels of difficulty, including tasks that are solvable by non-expert humans in seconds to extremely challenging ones that may take years to master (e.g., the NetHack Learning Environment). We devise fine-grained metrics to measure performance and conduct an extensive evaluation of several popular open-source and closed-source LLMs and VLMs. Our findings indicate that while current models achieve partial success in the easier games, they struggle significantly with more challenging tasks. Notably, we observe severe deficiencies in vision-based decision-making, as models perform worse when visual representations of the environments are provided. We release BALROG as an open and user-friendly benchmark to facilitate future research and development in the agentic community.

Metacognition--the awareness and regulation of one's cognitive processes--is central to human adaptability in unknown situations. In contrast, current autonomous agents often struggle in novel environments due to their limited capacity for adaptation. We hypothesize that metacognition is a critical missing ingredient in adaptive autonomous systems, equipping them with the cognitive flexibility needed to tackle unfamiliar challenges. Given the broad scope of metacognitive abilities, we focus on two key aspects: competence awareness and strategy selection for novel tasks. To this end, we propose the Metacognition for Unknown Situations and Environments (MUSE) framework, which integrates metacognitive processes--specifically self-awareness and self-regulation--into autonomous agents. We present two initial implementations of MUSE: one based on world modeling and another leveraging large language models (LLMs), both instantiating the metacognitive cycle. Our system continuously learns to assess its competence on a given task and uses this self-awareness to guide iterative cycles of strategy selection. MUSE agents show significant improvements in self-awareness and self-regulation, enabling them to solve novel, out-of-distribution tasks more effectively compared to Dreamer-v3-based reinforcement learning and purely prompt-based LLM agent approaches. This work highlights the promise of approaches inspired by cognitive and neural systems in enabling autonomous systems to adapt to new environments, overcoming the limitations of current methods that rely heavily on extensive training data.

The increasing demand for multilingual capabilities in healthcare underscores the need for AI models adept at processing diverse languages, particularly in clinical documentation and decision-making. Arabic, with its complex morphology, syntax, and diglossia, poses unique challenges for natural language processing (NLP) in medical contexts. This case study evaluates Sporo AraSum, a language model tailored for Arabic clinical documentation, against JAIS, the leading Arabic NLP model. Using synthetic datasets and modified PDQI-9 metrics modified ourselves for the purposes of assessing model performances in a different language. The study assessed the models' performance in summarizing patient-physician interactions, focusing on accuracy, comprehensiveness, clinical utility, and linguistic-cultural competence. Results indicate that Sporo AraSum significantly outperforms JAIS in AI-centric quantitative metrics and all qualitative attributes measured in our modified version of the PDQI-9. AraSum's architecture enables precise and culturally sensitive documentation, addressing the linguistic nuances of Arabic while mitigating risks of AI hallucinations. These findings suggest that Sporo AraSum is better suited to meet the demands of Arabic-speaking healthcare environments, offering a transformative solution for multilingual clinical workflows. Future research should incorporate real-world data to further validate these findings and explore broader integration into healthcare systems.

Large language models (LLMs) used across enterprises often use proprietary models and operate on sensitive inputs and data. The wide range of attack vectors identified in prior research - targeting various software and hardware components used in training and inference - makes it extremely challenging to enforce confidentiality and integrity policies. As we advance towards constructing compound AI inference pipelines that integrate multiple large language models (LLMs), the attack surfaces expand significantly. Attackers now focus on the AI algorithms as well as the software and hardware components associated with these systems. While current research often examines these elements in isolation, we find that combining cross-layer attack observations can enable powerful end-to-end attacks with minimal assumptions about the threat model. Given, the sheer number of existing attacks at each layer, we need a holistic and systemized understanding of different attack vectors at each layer. This SoK discusses different software and hardware attacks applicable to compound AI systems and demonstrates how combining multiple attack mechanisms can reduce the threat model assumptions required for an isolated attack. Next, we systematize the ML attacks in lines with the Mitre Att&ck framework to better position each attack based on the threat model. Finally, we outline the existing countermeasures for both software and hardware layers and discuss the necessity of a comprehensive defense strategy to enable the secure and high-performance deployment of compound AI systems.