This study presents the first comprehensive evaluation of Multimodal Large Language Models (MLLMs) for Adolescent Idiopathic Scoliosis (AIS) self-management. We constructed a database of approximately 3,000 anteroposterior X-rays with diagnostic texts and evaluated five MLLMs through a `Divide and Conquer' framework consisting of a visual question-answering task, a domain knowledge assessment task, and a patient education counseling assessment task. Our investigation revealed limitations of MLLMs' ability in interpreting complex spinal radiographs and comprehending AIS care knowledge. To address these, we pioneered enhancing MLLMs with spinal keypoint prompting and compiled an AIS knowledge base for retrieval augmented generation (RAG), respectively. Results showed varying effectiveness of visual prompting across different architectures, while RAG substantially improved models' performances on the knowledge assessment task. Our findings indicate current MLLMs are far from capable in realizing personalized assistant in AIS care. The greatest challenge lies in their abilities to obtain accurate detections of spinal deformity locations (best accuracy: 0.55) and directions (best accuracy: 0.13).

The reconstruction of visual information from brain activity fosters interdisciplinary integration between neuroscience and computer vision. However, existing methods still face challenges in accurately recovering highly complex visual stimuli. This difficulty stems from the characteristics of natural scenes: low-level features exhibit heterogeneity, while high-level features show semantic entanglement due to contextual overlaps. Inspired by the hierarchical representation theory of the visual cortex, we propose the HAVIR model, which separates the visual cortex into two hierarchical regions and extracts distinct features from each. Specifically, the Structural Generator extracts structural information from spatial processing voxels and converts it into latent diffusion priors, while the Semantic Extractor converts semantic processing voxels into CLIP embeddings. These components are integrated via the Versatile Diffusion model to synthesize the final image. Experimental results demonstrate that HAVIR enhances both the structural and semantic quality of reconstructions, even in complex scenes, and outperforms existing models.

Figure 1: TriVLA is a unified Vision-Language-Action framework that adopts a triple-system architecture inspired by the episodic world model. Image and language inputs are processed by a Vision-Language Model for multimodal perception. A Video Diffusion Model provides dynamic world modeling and future prediction. The policy module integrates sequential outputs, robot state, and action history and generates real-time actions for complex manipulation tasks. Recent advances in vision-language models (VLMs) have enabled robots to follow open-ended instructions and demonstrate impressive commonsense reasoning. However, current vision-language-action (VLA) frameworks primarily rely on static representations and limited temporal context, restricting agents to short-horizon, reactive behaviors and hindering robust generalization in dynamic embodied environments. Inspired by cognitive neuroscience theories of episodic memory, we are, to our knowledge, among the first to introduce a formalized episodic world model in VLA, enabling embodied robots to accumulate, recall, and predict sequential experiences. As an instantiation of this concept, our unified TriVLA realizes the episodic world model through a triple-system architecture: integrating multimodal grounding from a pretrained VLM (System 2) and temporally rich dynamics perception from a video diffusion model (System 3). This enables the agent to accumulate and recall sequential experiences, interpret current contexts, and predict future environmental evolution. Guided by episodic representations that span both the past and anticipated future, the downstream policy (System 1) generates coherent, context-aware action sequences through flow-matching and cross-modal attention mechanisms. It demonstrates strong long-horizon planning and open-ended intent understanding, showcasing the advantages of episodic world model-inspired reasoning for robust, generalizable robot intelligence. "Episodic memory is the only memory system that allows mental time travel--backward into the past and forward into the future. " -- Endel Tulving Building on this cognitive foundation, we advocate that robotic agents, require an internal episodic world model: a representational system that not only recalls past interactions but also anticipates future dynamics, thereby enabling robust generalization in embodied environments. Decades of cognitive neuroscience provide compelling evidence for this perspective.

Recent advances in large reasoning models (LRMs) have enabled strong chain-of-thought (CoT) generation through test-time computation. While these multi-step reasoning capabilities represent a major milestone in language model performance, they also introduce new safety risks. In this work, we present the first systematic study to revisit the problem of machine unlearning in the context of LRMs. Machine unlearning refers to the process of removing the influence of sensitive, harmful, or undesired data or knowledge from a trained model without full retraining. We show that conventional unlearning algorithms, originally designed for non-reasoning models, are inadequate for LRMs. In particular, even when final answers are successfully erased, sensitive information often persists within the intermediate reasoning steps, i.e., CoT trajectories. To address this challenge, we extend conventional unlearning and propose Reasoning-aware Representation Misdirection for Unlearning ($R^2MU$), a novel method that effectively suppresses sensitive reasoning traces and prevents the generation of associated final answers, while preserving the model's reasoning ability. Our experiments demonstrate that $R^2MU$ significantly reduces sensitive information leakage within reasoning traces and achieves strong performance across both safety and reasoning benchmarks, evaluated on state-of-the-art models such as DeepSeek-R1-Distill-LLaMA-8B and DeepSeek-R1-Distill-Qwen-14B.

Autoregressive language models (LMs) generate one token at a time, yet human reasoning operates over higher-level abstractions - sentences, propositions, and concepts. This contrast raises a central question- Can LMs likewise learn to reason over structured semantic units rather than raw token sequences? In this work, we investigate whether pretrained LMs can be lifted into such abstract reasoning spaces by building on their learned representations. We present a framework that adapts a pretrained token-level LM to operate in sentence space by autoregressively predicting continuous embeddings of next sentences. We explore two embedding paradigms inspired by classical representation learning: 1) semantic embeddings, learned via autoencoding to preserve surface meaning; and 2) contextual embeddings, trained via next-sentence prediction to encode anticipatory structure. We evaluate both under two inference regimes: Discretized, which decodes each predicted embedding into text before re-encoding; and Continuous, which reasons entirely in embedding space for improved efficiency. Across four domains - mathematics, logic, commonsense, and planning - contextual embeddings under continuous inference show competitive performance with Chain-of-Thought (CoT) while reducing inference-time FLOPs on average by half. We also present early signs of scalability and modular adaptation. Finally, to visualize latent trajectories, we introduce SentenceLens, a diagnostic tool that decodes intermediate model states into interpretable sentences. Together, our results indicate that pretrained LMs can effectively transition to abstract, structured reasoning within latent embedding spaces.

Large Language Models (LLMs) have recently demonstrated strong capabilities in code-related tasks, but their robustness in code reasoning under perturbations remains underexplored. We introduce CodeCrash, a stress-testing framework with 1,279 questions from CruxEval and LiveCodeBench, designed to evaluate reasoning reliability under structural perturbations and misleading natural language (NL) contexts. Through a systematic evaluation of 17 LLMs, we find that models often shortcut reasoning by over-relying on NL cues, leading to an average performance degradation of 23.2% in output prediction tasks. Even with Chain-of-Thought reasoning, models on average still have a 13.8% drop due to distractibility and rationalization, revealing a lack of critical reasoning capability to distinguish the actual code behaviors. While Large Reasoning Models with internal reasoning mechanisms improve robustness by fostering critical thinking, plausible yet incorrect hints can trigger pathological self-reflection, causing 2-3 times token consumption and even catastrophic cognitive dissonance in extreme cases for QwQ-32B. We refer to this phenomenon as Reasoning Collapse. CodeCrash provides a rigorous benchmark for evaluating robustness in code reasoning, guiding future research and development toward more reliable and resilient models.

Games have long been a microcosm for studying planning and reasoning in both natural and artificial intelligence, especially with a focus on expert-level or even super-human play. But real life also pushes human intelligence along a different frontier, requiring people to flexibly navigate decision-making problems that they have never thought about before. Here, we use novice gameplay to study how people make decisions and form judgments in new problem settings. We show that people are systematic and adaptively rational in how they play a game for the first time, or evaluate a game (e.g., how fair or how fun it is likely to be) before they have played it even once. We explain these capacities via a computational cognitive model that we call the "Intuitive Gamer". The model is based on mechanisms of fast and flat (depth-limited) goal-directed probabilistic simulation--analogous to those used in Monte Carlo tree-search models of expert game-play, but scaled down to use very few stochastic samples, simple goal heuristics for evaluating actions, and no deep search. In a series of large-scale behavioral studies with over 1000 participants and 121 two-player strategic board games (almost all novel to our participants), our model quantitatively captures human judgments and decisions varying the amount and kind of experience people have with a game--from no experience at all ("just thinking"), to a single round of play, to indirect experience watching another person and predicting how they should play--and does so significantly better than much more compute-intensive expert-level models. More broadly, our work offers new insights into how people rapidly evaluate, act, and make suggestions when encountering novel problems, and could inform the design of more flexible and human-like AI systems that can determine not just how to solve new tasks, but whether a task is worth thinking about at all.

Recent advancements in large multimodal models (LMMs) have shown strong capabilities in audio understanding. However, most systems rely solely on end-to-end reasoning, limiting interpretability and accuracy for tasks that require structured knowledge or specialized signal analysis. In this work, we present Audio-Maestro -- a tool-augmented audio reasoning framework that enables audio-language models to autonomously call external tools and integrate their timestamped outputs into the reasoning process. This design allows the model to analyze, transform, and interpret audio signals through specialized tools rather than relying solely on end-to-end inference. Experiments show that Audio-Maestro consistently improves general audio reasoning performance: Gemini-2.5-flash's average accuracy on MMAU-Test rises from 67.4% to 72.1%, DeSTA-2.5 from 58.3% to 62.8%, and GPT-4o from 60.8% to 63.9%. To our knowledge, Audio-Maestro is the first framework to integrate structured tool output into the large audio language model reasoning process.

Personalized alignment is crucial for enabling Large Language Models (LLMs) to engage effectively in user-centric interactions. However, current methods face a dual challenge: they fail to infer users' deep implicit preferences (including unstated goals, semantic context and risk tolerances), and they lack the defensive reasoning required to navigate real-world ambiguity. This cognitive gap leads to responses that are superficial, brittle and short-sighted. To address this, we propose Critique-Driven Reasoning Alignment (CDRA), which reframes alignment from a scalar reward-matching task into a structured reasoning process. First, to bridge the preference inference gap, we introduce the DeepPref benchmark. This dataset, comprising 3000 preference-query pairs across 20 topics, is curated by simulating a multi-faceted cognitive council that produces critique-annotated reasoning chains to deconstruct query semantics and reveal latent risks. Second, to instill defensive reasoning, we introduce the Personalized Generative Process Reward Model (Pers-GenPRM), which frames reward modeling as a personalized reasoning task. It generates a critique chain to evaluate a response's alignment with user preferences before outputting a final score based on this rationale. Ultimately, this interpretable, structured reward signal guides policy model through Critique-Driven Policy Alignment, a process-level online reinforcement learning algorithm integrating both numerical and natural language feedback. Experiments demonstrate that CDRA excels at discovering and aligning with users' true preferences while executing robust reasoning. Large Language Models (LLMs) are rapidly evolving from simple instruction-followers to potential collaborative partners, a transition that hinges on effective personalization (V aswani et al., 2023; Brown et al., 2020).

We present ViDRiP-LLaVA, the first large multimodal model (LMM) in computational pathology that integrates three distinct image scenarios, including single patch images, automatically segmented pathology video clips, and manually segmented pathology videos. This integration closely mirrors the natural diagnostic process of pathologists. By generating detailed histological descriptions and culminating in a definitive sign-out diagnosis, ViDRiP-LLaVA bridges visual narratives with diagnostic reasoning. Central to our approach is the ViDRiP-Instruct dataset, comprising 4278 video and diagnosis-specific chain-of-thought instructional pairs sourced from educational histopathology videos on YouTube. Although high-quality data is critical for enhancing diagnostic reasoning, its creation is time-intensive and limited in volume. To overcome this challenge, we transfer knowledge from existing single-image instruction datasets to train on weakly annotated, keyframe-extracted clips, followed by fine-tuning on manually segmented videos. ViDRiP-LLaVA establishes a new benchmark in pathology video analysis and offers a promising foundation for future AI systems that support clinical decision-making through integrated visual and diagnostic reasoning. Our code, data, and model are publicly available at: https://github.com/QuIIL/ViDRiP-LLaVA.