Kevin Zhu

Daniel Arnould, Rashad Aziz, Zixuan Kang et al.

Large Language Models (LLMs) excel at static reasoning tasks, yet their performance often degrades in interactive scenarios where information must be actively acquired through questioning. A key challenge lies in selecting questions that reduce uncertainty while incorporating responses that may be ambiguous or only partially informative. To address this, we propose Conversation-Aware Bayesian Experimental Design (CA-BED), an inference-time probabilistic dialog planning framework that integrates Bayesian Experimental Design with LLM-based likelihood estimation to optimize question selection over multiple conversational turns. CA-BED maintains a belief distribution over hypotheses, anticipates possible answers, and propagates expected information gain through a simulated conversation tree. Across two structured entity-deduction benchmarks, CA-BED yields an average 21.8% improvement in success rates over direct prompting, with comparable gains relative to alternative information-seeking methods. It achieves these gains with an average increase of only 1.8 conversational turns compared to direct prompting.

Daniel Flam-Shepherd, Kevin Zhu, Alán Aspuru-Guzik

Protein language models learn powerful representations directly from sequences of amino acids. However, they are constrained to generate proteins with only the set of amino acids represented in their vocabulary. In contrast, chemical language models learn atom-level representations of smaller molecules that include every atom, bond, and ring. In this work, we show that chemical language models can learn atom-level representations of proteins enabling protein generation unconstrained to the standard genetic code and far beyond it. In doing so, we show that language models can generate entire proteins atom by atom -- effectively learning the multiple hierarchical layers of molecular information that define proteins from their primary sequence to their secondary, and tertiary structure. We demonstrate language models are able to explore beyond protein space -- generating proteins with modified sidechains that form unnatural amino acids. Even further, we find that language models can explore chemical space and protein space simultaneously and generate novel examples of protein-drug conjugates. The results demonstrate the potential for biomolecular design at the atom level using language models.

Kevin Zhu, Alexander Wong, John McPhee

Current data analysis for the Canadian Olympic fencing team is primarily done manually by coaches and analysts. Due to the highly repetitive, yet dynamic and subtle movements in fencing, manual data analysis can be inefficient and inaccurate. We propose FenceNet as a novel architecture to automate the classification of fine-grained footwork techniques in fencing. FenceNet takes 2D pose data as input and classifies actions using a skeleton-based action recognition approach that incorporates temporal convolutional networks to capture temporal information. We train and evaluate FenceNet on the Fencing Footwork Dataset (FFD), which contains 10 fencers performing 6 different footwork actions for 10-11 repetitions each (652 total videos). FenceNet achieves 85.4% accuracy under 10-fold cross-validation, where each fencer is left out as the test set. This accuracy is within 1% of the current state-of-the-art method, JLJA (86.3%), which selects and fuses features engineered from skeleton data, depth videos, and inertial measurement units. BiFenceNet, a variant of FenceNet that captures the "bidirectionality" of human movement through two separate networks, achieves 87.6% accuracy, outperforming JLJA. Since neither FenceNet nor BiFenceNet requires data from wearable sensors, unlike JLJA, they could be directly applied to most fencing videos, using 2D pose data as input extracted from off-the-shelf 2D human pose estimators. In comparison to JLJA, our methods are also simpler as they do not require manual feature engineering, selection, or fusion.

Abhay Gupta, Philip Meng, Ece Yurtseven et al.

Detecting biases in natural language understanding (NLU) for African American Vernacular English (AAVE) is crucial to developing inclusive natural language processing (NLP) systems. To address dialect-induced performance discrepancies, we introduce AAVENUE ({AAVE} {N}atural Language {U}nderstanding {E}valuation), a benchmark for evaluating large language model (LLM) performance on NLU tasks in AAVE and Standard American English (SAE). AAVENUE builds upon and extends existing benchmarks like VALUE, replacing deterministic syntactic and morphological transformations with a more flexible methodology leveraging LLM-based translation with few-shot prompting, improving performance across our evaluation metrics when translating key tasks from the GLUE and SuperGLUE benchmarks. We compare AAVENUE and VALUE translations using five popular LLMs and a comprehensive set of metrics including fluency, BARTScore, quality, coherence, and understandability. Additionally, we recruit fluent AAVE speakers to validate our translations for authenticity. Our evaluations reveal that LLMs consistently perform better on SAE tasks than AAVE-translated versions, underscoring inherent biases and highlighting the need for more inclusive NLP models. We have open-sourced our source code on GitHub and created a website to showcase our work at https://aavenuee.github.io.

Samir Kassam, Angelo Markham, Katie Vo et al.

Gliomas, a common type of malignant brain tumor, present significant surgical challenges due to their similarity to healthy tissue. Preoperative Magnetic Resonance Imaging (MRI) images are often ineffective during surgery due to factors such as brain shift, which alters the position of brain structures and tumors. This makes real-time intraoperative MRI (ioMRI) crucial, as it provides updated imaging that accounts for these shifts, ensuring more accurate tumor localization and safer resections. This paper presents a deep learning pipeline combining You Only Look Once Version 8 (YOLOv8) and Segment Anything Model Vision Transformer-base (SAM ViT-b) to enhance glioma detection and segmentation during ioMRI. Our model was trained using the Brain Tumor Segmentation 2021 (BraTS 2021) dataset, which includes standard magnetic resonance imaging (MRI) images, and noise-augmented MRI images that simulate ioMRI images. Noised MRI images are harder for a deep learning pipeline to segment, but they are more representative of surgical conditions. Achieving a Dice Similarity Coefficient (DICE) score of 0.79, our model performs comparably to state-of-the-art segmentation models tested on noiseless data. This performance demonstrates the model's potential to assist surgeons in maximizing tumor resection and improving surgical outcomes.

Dharunish Yugeswardeenoo, Kevin Zhu, Sean O'Brien

Although LLMs have the potential to transform many fields, they still underperform humans in reasoning tasks. Existing methods induce the model to produce step-by-step calculations, but this research explores the question: Does making the LLM analyze the question improve its performance? We propose a novel prompting strategy called Question Analysis Prompting (QAP), in which the model is prompted to explain the question in $n$ words before solving. The value of $n$ influences the length of response generated by the model. QAP is evaluated on GPT 3.5 Turbo and GPT 4 Turbo on arithmetic datasets GSM8K, AQuA, and SAT and commonsense dataset StrategyQA. QAP is compared with other state-of-the-art prompts including Chain-of-Thought (CoT), Plan and Solve Prompting (PS+) and Take A Deep Breath (TADB). QAP outperforms all state-of-the-art prompts on AQuA and SAT datasets on both GPT3.5 and GPT4. QAP consistently ranks among the top-2 prompts on 75\% of the tests. A key factor of QAP performance can be attributed to response length, where detailed responses are beneficial when answering harder questions, but can negatively affect easy questions.

Aashray Reddy, Andrew Zagula, Nicholas Saban et al.

Large Language Models (LLMs) remain vulnerable to jailbreaking attacks where adversarial prompts elicit harmful outputs, yet most evaluations focus on single-turn interactions while real-world attacks unfold through adaptive multi-turn conversations. We present AutoAdv, a training-free framework for automated multi-turn jailbreaking that achieves up to 95% attack success rate on Llama-3.1-8B within six turns a 24 percent improvement over single turn baselines. AutoAdv uniquely combines three adaptive mechanisms: a pattern manager that learns from successful attacks to enhance future prompts, a temperature manager that dynamically adjusts sampling parameters based on failure modes, and a two-phase rewriting strategy that disguises harmful requests then iteratively refines them. Extensive evaluation across commercial and open-source models (GPT-4o-mini, Qwen3-235B, Mistral-7B) reveals persistent vulnerabilities in current safety mechanisms, with multi-turn attacks consistently outperforming single-turn approaches. These findings demonstrate that alignment strategies optimized for single-turn interactions fail to maintain robustness across extended conversations, highlighting an urgent need for multi-turn-aware defenses.

Joongho Kim, Xirui Huang, Zarreen Reza et al.

Tree-of-Thought (ToT) reasoning boosts the problem-solving abilities of Large Language Models (LLMs) but is computationally expensive due to semantic redundancy, where distinct branches explore equivalent reasoning paths. We introduce Semantic Similarity-Based Dynamic Pruning (SSDP), a lightweight method that, to the best of our knowledge, is the first framework to integrate online semantic merging into parallelized tree search, enabling the clustering and pruning of redundant steps in real time. Across reasoning benchmarks, including GSM8K and MATH500, SSDP achieves up to a 2.3x speedup over state-of-the-art tree-search baselines while maintaining competitive accuracy (typically within 5% of the strongest baseline) and reducing the number of explored nodes by 85-90%, demonstrating a practical approach to efficient, scalable LLM reasoning. The implementation of SSDP is publicly available at https://github.com/kimjoonghokim/SSDP.

Rohan Subramanian Thomas, Shikhar Shiromani, Abdullah Chaudhry et al.

Prompt design significantly impacts the moral competence and safety alignment of large language models (LLMs), yet empirical comparisons remain fragmented across datasets and models.We introduce ProMoral-Bench, a unified benchmark evaluating 11 prompting paradigms across four LLM families. Using ETHICS, Scruples, WildJailbreak, and our new robustness test, ETHICS-Contrast, we measure performance via our proposed Unified Moral Safety Score (UMSS), a metric balancing accuracy and safety. Our results show that compact, exemplar-guided scaffolds outperform complex multi-stage reasoning, providing higher UMSS scores and greater robustness at a lower token cost. While multi-turn reasoning proves fragile under perturbations, few-shot exemplars consistently enhance moral stability and jailbreak resistance. ProMoral-Bench establishes a standardized framework for principled, cost-effective prompt engineering.

Kaylee Chhua, Zhoujinyi Wen, Vedant Hathalia et al.

This study addresses the racial biases in facial expression recognition (FER) systems within Large Multimodal Foundation Models (LMFMs). Despite advances in deep learning and the availability of diverse datasets, FER systems often exhibit higher error rates for individuals with darker skin tones. Existing research predominantly focuses on traditional FER models (CNNs, RNNs, ViTs), leaving a gap in understanding racial biases in LMFMs. We benchmark four leading LMFMs: GPT-4o, PaliGemma, Gemini, and CLIP to assess their performance in facial emotion detection across different racial demographics. A linear classifier trained on CLIP embeddings obtains accuracies of 95.9\% for RADIATE, 90.3\% for Tarr, and 99.5\% for Chicago Face. Furthermore, we identify that Anger is misclassified as Disgust 2.1 times more often in Black Females than White Females. This study highlights the need for fairer FER systems and establishes a foundation for developing unbiased, accurate FER technologies. Visit https://kvjvhub.github.io/FERRacialBias/ for further information regarding the biases within facial expression recognition.

Daniel Csizmadia, Andrei Codreanu, Victor Sim et al.

We present Distill CLIP (DCLIP), a fine-tuned variant of the CLIP model that enhances multimodal image-text retrieval while preserving the original model's strong zero-shot classification capabilities. CLIP models are typically constrained by fixed image resolutions and limited context, which can hinder their effectiveness in retrieval tasks that require fine-grained cross-modal understanding. DCLIP addresses these challenges through a meta teacher-student distillation framework, where a cross-modal transformer teacher is fine-tuned to produce enriched embeddings via bidirectional cross-attention between YOLO-extracted image regions and corresponding textual spans. These semantically and spatially aligned global representations guide the training of a lightweight student model using a hybrid loss that combines contrastive learning and cosine similarity objectives. Despite being trained on only ~67,500 samples curated from MSCOCO, Flickr30k, and Conceptual Captions-just a fraction of CLIP's original dataset-DCLIP significantly improves image-text retrieval metrics (Recall@K, MAP), while retaining approximately 94% of CLIP's zero-shot classification performance. These results demonstrate that DCLIP effectively mitigates the trade-off between task specialization and generalization, offering a resource-efficient, domain-adaptive, and detail-sensitive solution for advanced vision-language tasks. Code available at https://anonymous.4open.science/r/DCLIP-B772/README.md.

Hanning Chen, Keyu Man, Kevin Zhu et al.

Identifying and addressing performance anti-patterns in machine learning (ML) models is critical for efficient training and inference, but it typically demands deep expertise spanning system infrastructure, ML models and kernel development. While large tech companies rely on dedicated ML infrastructure engineers to analyze torch traces and benchmarks, such resource-intensive workflows are largely inaccessible to computer vision researchers in general. Among the challenges, pinpointing problematic trace segments within lengthy execution traces remains the most time-consuming task, and is difficult to automate with current ML models, including LLMs. In this work, we present the first benchmark dataset specifically designed to evaluate and improve ML models' ability to detect anti patterns in traces. Our dataset contains over 600 PyTorch traces from diverse computer vision models classification, detection, segmentation, and generation collected across multiple hardware platforms. We also propose a novel iterative approach: a lightweight ML model first detects trace segments with anti patterns, followed by a large language model (LLM) for fine grained classification and targeted feedback. Experimental results demonstrate that our method significantly outperforms unsupervised clustering and rule based statistical techniques for detecting anti pattern regions. Our method also effectively compensates LLM's limited context length and reasoning inefficiencies.

Sundesh Donthi, Maximilian Spencer, Om Patel et al.

For large language models (LLMs) like NLLB and GPT, translating idioms remains a challenge. Our goal is to enhance translation fidelity by improving LLM processing of idiomatic language while preserving the original linguistic style. This has a significant social impact, as it preserves cultural nuances and ensures translated texts retain their intent and emotional resonance, fostering better cross-cultural communication. Previous work has utilized knowledge bases like IdiomKB by providing the LLM with the meaning of an idiom to use in translation. Although this method yielded better results than a direct translation, it is still limited in its ability to preserve idiomatic writing style across languages. In this research, we expand upon the knowledge base to find corresponding idioms in the target language. Our research performs translations using two methods: The first method employs the SentenceTransformers model to semantically generate cosine similarity scores between the meanings of the original and target language idioms, selecting the best idiom (Cosine Similarity method). The second method uses an LLM to find a corresponding idiom in the target language for use in the translation (LLM-generated idiom method). As a baseline, we performed a direct translation without providing additional information. Human evaluations on the English -> Chinese, and Chinese -> English show the Cosine Similarity Lookup method out-performed others in all GPT4o translations. To further build upon IdiomKB, we developed a low-resource Urdu dataset containing Urdu idioms and their translations. Despite dataset limitations, the Cosine Similarity Lookup method shows promise, potentially overcoming language barriers and enabling the exploration of diverse literary works in Chinese and Urdu.(LoResLM @ COLING Preprint)

Anand Swaroop, Akshat Nallani, Saksham Uboweja et al.

Chain-of-thought (CoT) reasoning has emerged as a powerful tool for improving large language model performance on complex tasks, but recent work shows that reasoning steps often fail to causally influence the final answer, creating brittle and untrustworthy outputs. Prior approaches focus primarily on measuring faithfulness, while methods for systematically improving it remain limited. We introduce Faithful Reasoning via Intervention Training (FRIT), a scalable alignment method that trains models to produce causally consistent reasoning by learning from systematically corrupted examples. FRIT generates synthetic training data by intervening on individual reasoning steps in model-generated CoTs, creating faithful/unfaithful pairs that highlight when reasoning breaks down. We then apply Direct Preference Optimization to teach models to prefer causally consistent reasoning paths. Evaluating on Qwen3-8B and Mistral-7B-v0.1 across factual and symbolic reasoning tasks, FRIT increases faithful reasoning by $3.4$ percentage points for Mistral on GSM8K while improving accuracy by $7.6$ percentage points. Our approach provides the first scalable, supervision-free method for training language models to produce more reliable and interpretable reasoning, addressing a critical gap between reasoning performance and trustworthiness. We release our code at \href{https://github.com/Anut-py/frit}.

Athena Wen, Tanush Patil, Ansh Saxena et al.

In an era where AI-driven hiring is transforming recruitment practices, concerns about fairness and bias have become increasingly important. To explore these issues, we introduce a benchmark, FAIRE (Fairness Assessment In Resume Evaluation), to test for racial and gender bias in large language models (LLMs) used to evaluate resumes across different industries. We use two methods-direct scoring and ranking-to measure how model performance changes when resumes are slightly altered to reflect different racial or gender identities. Our findings reveal that while every model exhibits some degree of bias, the magnitude and direction vary considerably. This benchmark provides a clear way to examine these differences and offers valuable insights into the fairness of AI-based hiring tools. It highlights the urgent need for strategies to reduce bias in AI-driven recruitment. Our benchmark code and dataset are open-sourced at our repository: https://github.com/athenawen/FAIRE-Fairness-Assessment-In-Resume-Evaluation.git.

Yixuan Yang, Mehak Arora, Ryan Zhang et al.

We present Clin-JEPA, a multi-phase co-training framework for joint-embedding predictive (JEPA) pretraining on EHR patient trajectories. JEPA architectures have enabled latent-space planning in robotics and high-quality representation learning in vision, but extending the paradigm to EHR data -- to obtain a single backbone that simultaneously forecasts patient trajectories and serves diverse downstream risk-prediction tasks without per-task fine-tuning -- remains an open challenge. Existing JEPA frameworks either discard the predictor after pretraining (I-JEPA, V-JEPA) or train it on a frozen pretrained encoder (V-JEPA 2-AC), leaving the encoder unaware of the rollout signal that the retained predictor must use at inference; co-training the encoder and predictor under a shared JEPA prediction objective would supply this grounding, but naïve co-training is unstable, with representation collapse and online/target drift causing autoregressive rollout to diverge. Clin-JEPA's five-phase pretraining curriculum -- predictor warmup, joint refinement, EMA target alignment, hard sync, and predictor finalization -- addresses each failure mode by phase, stably co-training a Qwen3-8B-based encoder and a 92M-parameter latent trajectory predictor. On MIMIC-IV ICU data, three independent evaluations support the framework: (1) latent $\ell_1$ rollout drift uniquely converges ($-$15.7%) over 48-hour horizons while baselines and ablations diverge (+3% to +4951%); (2) the encoder learns a clinically discriminative latent geometry (deteriorating-patient cohorts displace 4.83$\times$ further than stable patients in latent space, vs $\leq$2.62$\times$ for baseline encoders); (3) a single backbone outperforms strong tabular and sequence baselines on multi-task downstream evaluation. Clin-JEPA achieves mean AUROC 0.851 on ICareFM EEP and 0.883 on 8 binary risk tasks (+0.038 and +0.041 vs baseline average).

Kevin Zhu, William Tang, Raphael Hay Tene et al.

Rapid and accurate damage assessment following natural disasters is critical for effective emergency response. However, identifying fine-grained damage levels (e.g., distinguishing minor from major roof damage) in UAV imagery remains challenging due to the degradation of texture cues during resizing and extreme class imbalance. We propose DA-SegFormer, a damage-aware adaptation of the SegFormer architecture optimized for high-resolution disaster imagery. Our method introduces a Class-Aware Sampling strategy to guarantee exposure to rare damage features, and it integrates Online Hard Example Mining (OHEM) with Dice Loss to dynamically focus on underrepresented classes. In addition, we employ a resolution-preserving inference protocol that maintains native texture details. Evaluated on the RescueNet dataset, DA-SegFormer achieves 74.61\% mIoU, outperforming the baseline by 2.55\%. Notably, our improvements yield double-digit gains in critical damage classes: Minor Damage (+11.7%) and Major Damage (+21.3%).

Rajat Rawat, Hudson McBride, Dhiyaan Nirmal et al.

As large language models (LLMs) gain traction in healthcare, concerns about their susceptibility to demographic biases are growing. We introduce {DiversityMedQA}, a novel benchmark designed to assess LLM responses to medical queries across diverse patient demographics, such as gender and ethnicity. By perturbing questions from the MedQA dataset, which comprises medical board exam questions, we created a benchmark that captures the nuanced differences in medical diagnosis across varying patient profiles. Our findings reveal notable discrepancies in model performance when tested against these demographic variations. Furthermore, to ensure the perturbations were accurate, we also propose a filtering strategy that validates each perturbation. By releasing DiversityMedQA, we provide a resource for evaluating and mitigating demographic bias in LLM medical diagnoses.

Elysia Shi, Adithri Manda, London Chowdhury et al.

When using AI to detect signs of depressive disorder, AI models habitually draw preemptive conclusions. We theorize that using chain-of-thought (CoT) prompting to evaluate Patient Health Questionnaire-8 (PHQ-8) scores will improve the accuracy of the scores determined by AI models. In our findings, when the models reasoned with CoT, the estimated PHQ-8 scores were consistently closer on average to the accepted true scores reported by each participant compared to when not using CoT. Our goal is to expand upon AI models' understanding of the intricacies of human conversation, allowing them to more effectively assess a patient's feelings and tone, therefore being able to more accurately discern mental disorder symptoms; ultimately, we hope to augment AI models' abilities, so that they can be widely accessible and used in the medical field.

Michael S. Zhang, Rishi A. Ruia, Arnav Kewalram et al.

Catastrophic forgetting poses a fundamental challenge in continual learning, particularly when models are quantized for deployment efficiency. We systematically investigate the interplay between quantization precision (FP16, INT8, INT4) and replay buffer strategies in large language models, revealing unexpected dynamics. While FP16 achieves superior initial task performance (74.44% on NLU), we observe a striking inversion on subsequent tasks: quantized models outperform FP16 by 8-15% on final task forward accuracy, with INT4 achieving nearly double FP16's performance on Code generation (40% vs 20%). Critically, even minimal replay buffers (0.1%) dramatically improve retention - increasing NLU retention after Math training from 45% to 65% across all precision levels - with INT8 consistently achieving the optimal balance between learning plasticity and knowledge retention. We hypothesize that quantization-induced noise acts as implicit regularization, preventing the overfitting to new task gradients that plagues high-precision models. These findings challenge the conventional wisdom that higher precision is always preferable, suggesting instead that INT8 quantization offers both computational efficiency and superior continual learning dynamics. Our results provide practical guidelines for deploying compressed models in continual learning scenarios: small replay buffers (1-2%) suffice for NLU tasks, while Math and Code benefit from moderate buffers (5-10%), with quantized models requiring less replay than FP16 to achieve comparable retention. Code is available at https://github.com/Festyve/LessIsMore.

Glenn Zhang, Treasure Mayowa, Jason Fan et al.

Producing trustworthy and reliable Large Language Models (LLMs) has become increasingly important as their usage becomes more widespread. Calibration seeks to achieve this by improving the alignment between the model's confidence and the actual likelihood of its responses being correct or desirable. However, it has been observed that the internal confidence of a model, derived from token probabilities, is not well aligned with its verbalized confidence, leading to misleading results with different calibration methods. In this paper, we propose Direct Confidence Alignment (DCA), a method using Direct Preference Optimization to align an LLM's verbalized confidence with its internal confidence rather than ground-truth accuracy, enhancing model transparency and reliability by ensuring closer alignment between the two confidence measures. We evaluate DCA across multiple open-weight LLMs on a wide range of datasets. To further assess this alignment, we also introduce three new calibration error-based metrics. Our results show that DCA improves alignment metrics on certain model architectures, reducing inconsistencies in a model's confidence expression. However, we also show that it can be ineffective on others, highlighting the need for more model-aware approaches in the pursuit of more interpretable and trustworthy LLMs.

Miguel Tjia, Artem Kim, Elaine Wynette Wijaya et al.

7,651 cases of Search and Rescue Missions (SAR) were reported by the United States Coast Guard in 2024, with over 1322 SAR helicopters deployed in the 6 first months alone. Through the utilizations of YOLO, we were able to run different weather conditions and lighting from our augmented dataset for training. YOLO then utilizes CNNs to apply a series of convolutions and pooling layers to the input image, where the convolution layers are able to extract the main features of the image. Through this, our YOLO model is able to learn to differentiate different objects which may considerably improve its accuracy, possibly enhancing the efficiency of SAR operations through enhanced detection accuracy. This paper aims to improve the model's accuracy of human detection in maritime SAR by evaluating a robust datasets containing various elevations and geological locations, as well as through data augmentation which simulates different weather and lighting. We observed that models trained on augmented datasets outperformed their non-augmented counterparts in which the human recall scores ranged from 0.891 to 0.911 with an improvement rate of 3.4\% on the YOLOv5l model. Results showed that these models demonstrate greater robustness to real-world conditions in varying of weather, brightness, tint, and contrast.

Dharunish Yugeswardeenoo, Harshil Nukala, Ved Shah et al.

Large Language Models (LLMs) have demonstrated impressive reasoning capabilities but continue to struggle with arithmetic tasks. Prior works largely focus on outputs or prompting strategies, leaving the open question of the internal structure through which models do arithmetic computation. In this work, we investigate whether LLMs encode operator precedence in their internal representations via the open-source instruction-tuned LLaMA 3.2-3B model. We constructed a dataset of arithmetic expressions with three operands and two operators, varying the order and placement of parentheses. Using this dataset, we trace whether intermediate results appear in the residual stream of the instruction-tuned LLaMA 3.2-3B model. We apply interpretability techniques such as logit lens, linear classification probes, and UMAP geometric visualization. Our results show that intermediate computations are present in the residual stream, particularly after MLP blocks. We also find that the model linearly encodes precedence in each operator's embeddings post attention layer. We introduce partial embedding swap, a technique that modifies operator precedence by exchanging high-impact embedding dimensions between operators.

Kiran Nijjer, Ryan Bui, Derek Jiu et al.

SkinGPT-4, a large vision-language model, leverages annotated skin disease images to augment clinical workflows in underserved communities. However, its training dataset predominantly represents lighter skin tones, limiting diagnostic accuracy for darker tones. Here, we evaluated performance biases in SkinGPT-4 across skin tones on common skin diseases, including eczema, allergic-contact dermatitis, and psoriasis using the open-sourced SCIN dataset. We leveraged the SkinGPT-4 backbone to develop finetuned models for custom skin disease classification tasks and explored bias mitigation strategies. Clinical evaluation by board-certified dermatologists on six relevant skin diseases from 300 SCIN cases assessed images for diagnostic accuracy, informativity, physician utility, and patient utility. Model fairness metrics, including demographic parity and equalized odds, were calculated across skin tones. SkinGPT-4 achieved an average demographic parity of 0.10 across Fitzpatrick types, with notable differences of 0.10-0.15 between lightest and darkest tones across evaluation metrics. Model hallucinations in artifacts and anatomy occurred at a rate of 17.8. Our customized models achieved average F1, precision, and AUROC of 0.75, 0.78, and 0.78 across visually similar disease pairs. Fairness analysis showed an average demographic parity of 0.75, with a maximum disparity of 0.21 across skin tones. The best model achieved parity scores of 0.83, 0.83, 0.76, 0.89, 0.90, and 0.90 for Fitzpatrick I-VI, indicating robust fairness. Large language models such as SkinGPT-4 showed weaker performance on darker tones. Model biases exist across evaluation criteria, and hallucinations may affect diagnostic efficacy. These findings demonstrate the efficacy of training accurate, fair models using existing backbones for custom skin disease classification.

Prameshwar Thiyagarajan, Vaishnavi Parimi, Shamant Sai et al.

Theory of Mind (ToM), the ability to understand the mental states of oneself and others, remains a challenging area for large language models (LLMs), which often fail to predict human mental states accurately. In this paper, we introduce UniToMBench, a unified benchmark that integrates the strengths of SimToM and TOMBENCH to systematically improve and assess ToM capabilities in LLMs by integrating multi-interaction task designs and evolving story scenarios. Supported by a custom dataset of over 1,000 hand-written scenarios, UniToMBench combines perspective-taking techniques with diverse evaluation metrics to better stimulate social cognition in LLMs. Through evaluation, we observe that while models like GPT-4o and GPT-4o Mini show consistently high accuracy in tasks involving emotional and belief-related scenarios, with results usually above 80%, there is significant variability in their performance across knowledge-based tasks. These results highlight both the strengths and limitations of current LLMs in ToM-related tasks, underscoring the value of UniToMBench as a comprehensive tool for future development. Our code is publicly available here: https://github.com/Shamant/unifiedtombenchmark.

Vijay Goyal, Mustafa Khan, Aprameya Tirupati et al.

Large language models (LLMs) have demonstrated remarkable performance across a wide range of natural language processing (NLP) tasks. However, these models are often difficult to deploy due to significant computational requirements and resource constraints. Knowledge distillation (KD) is an effective technique for transferring the performance of larger LLMs to smaller models. Traditional KD methods primarily focus on the direct output of the teacher model, with little emphasis on the role of prompting during knowledge transfer. In this paper, we propose a set of novel response-priming prompting strategies applied in the knowledge distillation pipeline to enhance the performance of student models. Our approach fine-tunes a smaller Llama 3.1 8B Instruct model by distilling knowledge from a quantized Llama 3.1 405B Instruct teacher model. We apply LoRA optimization and evaluate on the GSM8K benchmark. Experimental results demonstrate that integrating reasoning-eliciting prompting into the proposed KD pipeline significantly improves student model performance, offering an efficient way to deploy powerful models in resource-constrained environments. We find that Ground Truth prompting results in a 55\% performance increase on GSM8K for a distilled Llama 3.1 8B Instruct compared to the same model distilled without prompting. A thorough investigation into the self-attention layers of the student models indicates that the more successful prompted models tend to exhibit certain positive behaviors inside their attention heads which can be tied to their increased accuracy. Our implementation can be found at https://github.com/alonso130r/knowledge-distillation.

Daniel Aarao Reis Arturi, Eric Zhang, Andrew Ansah et al.

Recent work has discovered that large language models can develop broadly misaligned behaviors after being fine-tuned on narrowly harmful datasets, a phenomenon known as emergent misalignment (EM). However, the fundamental mechanisms enabling such harmful generalization across disparate domains remain poorly understood. In this work, we adopt a geometric perspective to study EM and demonstrate that it exhibits a fundamental cross-task linear structure in how harmful behavior is encoded across different datasets. Specifically, we find a strong convergence in EM parameters across tasks, with the fine-tuned weight updates showing relatively high cosine similarities, as well as shared lower-dimensional subspaces as measured by their principal angles and projection overlaps. Furthermore, we also show functional equivalence via linear mode connectivity, wherein interpolated models across narrow misalignment tasks maintain coherent, broadly misaligned behavior. Our results indicate that EM arises from different narrow tasks discovering the same set of shared parameter directions, suggesting that harmful behaviors may be organized into specific, predictable regions of the weight landscape. By revealing this fundamental connection between parametric geometry and behavioral outcomes, we hope our work catalyzes further research on parameter space interpretability and weight-based interventions.

Jay Shim, Grant Kruttschnitt, Alyssa Ma et al.

Rapidly increasing model scales coupled with steering methods such as chain-of-thought prompting have led to drastic improvements in language model reasoning. At the same time, models struggle with compositional generalization and are far from human performance on many reasoning-based benchmarks. Leveraging the success of chain-of-thought prompting, and also taking inspiration from context-aware decoding (CAD), we explore input-based contrasting methods to further encourage the type of reasoning induced by chain-of-thought prompting. While work remains to stabilize these results across datasets and models, the improvements we find warrant further investigation into input-based steering methods for context-aware reasoning.

Edward Lue Chee Lip, Anthony Channg, Diana Kim et al.

As AI capabilities advance, we increasingly rely on powerful models to decompose complex tasks $\unicode{x2013}$ but what if the decomposer itself is malicious? Factored cognition protocols decompose complex tasks into simpler child tasks: one model creates the decomposition, while other models implement the child tasks in isolation. Prior work uses trusted (weaker but reliable) models for decomposition, which limits usefulness for tasks where decomposition itself is challenging. We introduce Factor($U$,$T$), in which an untrusted (stronger but potentially malicious) model decomposes while trusted models implement child tasks. Can monitors detect malicious activity when observing only natural language task instructions, rather than complete solutions? We baseline and red team Factor($U$,$T$) in control evaluations on BigCodeBench, a dataset of Python coding tasks. Monitors distinguishing malicious from honest decompositions perform poorly (AUROC 0.52) compared to monitors evaluating complete Python solutions (AUROC 0.96). Furthermore, Factor($D$,$U$), which uses a trusted decomposer and monitors concrete child solutions, achieves excellent discrimination (AUROC 0.96) and strong safety (1.2% ASR), demonstrating that implementation-context monitoring succeeds where decomposition-only monitoring fails.

Pascal Passigan, Kevin Zhu, Angelina Ning

Understanding how small molecules perturb gene expression is essential for uncovering drug mechanisms, predicting off-target effects, and identifying repurposing opportunities. While prior deep learning frameworks have integrated multimodal embeddings into biomedical knowledge graphs (BKGs) and further improved these representations through graph neural network message-passing paradigms, these models have been applied to tasks such as link prediction and binary drug-disease association, rather than the task of gene perturbation, which may unveil more about mechanistic transcriptomic effects. To address this gap, we construct a merged biomedical graph that integrates (i) PrimeKG++, an augmentation of PrimeKG containing semantically rich embeddings for nodes with (ii) LINCS L1000 drug and cell line nodes, initialized with multimodal embeddings from foundation models such as MolFormerXL and BioBERT. Using this heterogeneous graph, we train a graph attention network (GAT) with a downstream prediction head that learns the delta expression profile of over 978 landmark genes for a given drug-cell pair. Our results show that our framework outperforms MLP baselines for differentially expressed genes (DEG) -- which predict the delta expression given a concatenated embedding of drug features, target features, and baseline cell expression -- under the scaffold and random splits. Ablation experiments with edge shuffling and node feature randomization further demonstrate that the edges provided by biomedical KGs enhance perturbation-level prediction. More broadly, our framework provides a path toward mechanistic drug modeling: moving beyond binary drug-disease association tasks to granular transcriptional effects of therapeutic intervention.

Anika Singh, Aayush Dhaulakhandi, Varun Chopade et al.

As neural networks are increasingly deployed in dynamic environments, they face the challenge of catastrophic forgetting, the tendency to overwrite previously learned knowledge when adapting to new tasks, resulting in severe performance degradation on earlier tasks. We propose Selective Forgetting-Aware Optimization (SFAO), a dynamic method that regulates gradient directions via cosine similarity and per-layer gating, enabling controlled forgetting while balancing plasticity and stability. SFAO selectively projects, accepts, or discards updates using a tunable mechanism with efficient Monte Carlo approximation. Experiments on standard continual learning benchmarks show that SFAO achieves competitive accuracy with markedly lower memory cost, a 90$\%$ reduction, and improved forgetting on MNIST datasets, making it suitable for resource-constrained scenarios.

Adam Kamel, Tanish Rastogi, Michael Ma et al.

Transformer-based large language models (LLMs) have demonstrated strong reasoning abilities across diverse fields, from solving programming challenges to competing in strategy-intensive games such as chess. Prior work has shown that LLMs can develop emergent world models in games of perfect information, where internal representations correspond to latent states of the environment. In this paper, we extend this line of investigation to domains of incomplete information, focusing on poker as a canonical partially observable Markov decision process (POMDP). We pretrain a GPT-style model on Poker Hand History (PHH) data and probe its internal activations. Our results demonstrate that the model learns both deterministic structure, such as hand ranks, and stochastic features, such as equity, without explicit instruction. Furthermore, by using primarily nonlinear probes, we demonstrated that these representations are decodeable and correlate with theoretical belief states, suggesting that LLMs are learning their own representation of the stochastic environment of Texas Hold'em Poker.

Rohin Sood, Kevin Zhu

Effective water quality monitoring in coastal regions is crucial due to the progressive deterioration caused by pollution and human activities. To address this, this study develops time-series models to predict chlorophyll-a (Chl-a), suspended solids (SS), and turbidity using Sentinel-2 satellite data and Google Earth Engine (GEE) in the coastal regions of Hong Kong. Leveraging Long Short-Term Memory (LSTM) Recurrent Neural Networks, the study incorporates extensive temporal datasets to enhance prediction accuracy. The models utilize spectral data from Sentinel-2, focusing on optically active components, and demonstrate that selected variables closely align with the spectral characteristics of Chl-a and SS. The results indicate improved predictive performance over previous methods, highlighting the potential for remote sensing technology in continuous and comprehensive water quality assessment.

Darsh Kachroo, Adriana Caraeni, Arjun Prasaath Anbazhagan et al.

Direct Preference Optimization (DPO) is an effective framework for aligning large language models with human preferences, but it struggles with complex reasoning tasks. DPO optimizes for the likelihood of generating preferred over dispreferred responses in their entirety and lacks the granularity to provide feedback on subsections of many-step solutions typical of reasoning tasks. Existing methods excel at either stable preference learning (e.g., DPO variants like KTO and RSO) or structured reasoning (e.g., ReMA's multi-agent RL framework, Tree of Thoughts), but fail to merge these complementary strengths. We propose HiPO (Hierarchical Preference Optimization), an extension of DPO that separates responses into reasoning segments (query clarification and context, reasoning steps, and answer) and computes loss as a weighted sum of the DPO loss for each segment. Our approach enables segment-specific training while maintaining DPO's computational efficiency and training stability. We demonstrate that for multiple 7B LLMs fine-tuned using HiPO and DPO on the Math Stack Exchange preference dataset, the models trained with HiPO outperform the others on a variety of common math benchmarks and achieve greater organization, logical flow, and consistency as measured by GPT-4.1.

Mrinal Agarwal, Saad Rana, Theo Sundoro et al.

Deception is a fundamental challenge for multi-agent reasoning: effective systems must strategically conceal information while detecting misleading behavior in others. Yet most evaluations reduce deception to static classification, ignoring the interactive, adversarial, and longitudinal nature of real deceptive dynamics. Large language models (LLMs) can deceive convincingly but remain weak at detecting deception in peers. We present WOLF, a multi-agent social deduction benchmark based on Werewolf that enables separable measurement of deception production and detection. WOLF embeds role-grounded agents (Villager, Werewolf, Seer, Doctor) in a programmable LangGraph state machine with strict night-day cycles, debate turns, and majority voting. Every statement is a distinct analysis unit, with self-assessed honesty from speakers and peer-rated deceptiveness from others. Deception is categorized via a standardized taxonomy (omission, distortion, fabrication, misdirection), while suspicion scores are longitudinally smoothed to capture both immediate judgments and evolving trust dynamics. Structured logs preserve prompts, outputs, and state transitions for full reproducibility. Across 7,320 statements and 100 runs, Werewolves produce deceptive statements in 31% of turns, while peer detection achieves 71-73% precision with ~52% overall accuracy. Precision is higher for identifying Werewolves, though false positives occur against Villagers. Suspicion toward Werewolves rises from ~52% to over 60% across rounds, while suspicion toward Villagers and the Doctor stabilizes near 44-46%. This divergence shows that extended interaction improves recall against liars without compounding errors against truthful roles. WOLF moves deception evaluation beyond static datasets, offering a dynamic, controlled testbed for measuring deceptive and detective capacity in adversarial multi-agent interaction.

Ian Su, Gaurav Purushothaman, Jey Narayan et al.

Reasoning-specialized models like OpenAI's 5.1 and DeepSeek-V3.2 allocate substantial inference compute to extended chain-of-thought (CoT) traces, yet reasoning tokens incur significant costs. How do different reasoning modalities of code, natural language, hybrid, or none do perform under token constraints? We introduce a framework that constrains models to reason exclusively through code, comments, both, or neither, then systematically ablates token budgets to 10\%, 30\%, 50\%, and 70\% of optimal. We evaluate four frontier models (GPT-5.1, Gemini 3 Flash, DeepSeek-V3.2, Grok 4.1) across mathematical benchmarks (AIME, GSM8K, HMMT). Our findings reveal: (1) \textbf{truncated reasoning can hurt} as DeepSeek-V3.2 achieves 53\% with no reasoning but only 17\% with truncated CoT at 50\% budget; (2) \textbf{code degrades gracefully} as Gemini's comments collapse to 0\% while code maintains 43-47\%; (3) \textbf{hybrid reasoning underperforms} single modalities; (4) \textbf{robustness is model-dependent} as Grok maintains 80-90\% at 30\% budget where OpenAI and DeepSeek collapse to 7-27\%. These results suggest incomplete reasoning chains actively mislead models, with implications for deploying reasoning-specialized systems under resource constraints.

Ben Zelenskiy, Saibilila Abudukelimu, George Flint et al.

We introduce the Sumudu Neural Operator (SNO), a neural operator rooted in the properties of the Sumudu Transform. We leverage the relationship between the polynomial expansions of transform pairs to decompose the input space as coefficients, which are then transformed into the Sumudu Space, where the neural operator is parameterized. We evaluate the operator in ODEs (Duffing Oscillator, Lorenz System, and Driven Pendulum) and PDEs (Euler-Bernoulli Beam, Burger's Equation, Diffusion, Diffusion-Reaction, and Brusselator). SNO achieves superior performance to FNO on PDEs and demonstrates competitive accuracy with LNO on several PDE tasks, including the lowest error on the Euler-Bernoulli Beam and Diffusion Equation. Additionally, we apply zero-shot super-resolution to the PDE tasks to observe the model's capability of obtaining higher quality data from low-quality samples. These preliminary findings suggest promise for the Sumudu Transform as a neural operator design, particularly for certain classes of PDEs.

Jiahang He, Rishi Ramachandran, Neel Ramachandran et al.

As large language models (LLMs) are adopted in an increasingly wide range of applications, user-model interactions have grown in both frequency and scale. Consequently, research has focused on evaluating the robustness of LLMs, an essential quality for real-world tasks. In this paper, we employ simple multi-turn follow-up prompts to evaluate models' answer changes, model accuracy dynamics across turns with Markov chains, and examine whether linear probes can predict these changes. Our results show significant vulnerabilities in LLM robustness: a simple "Think again" prompt led to an approximate 10% accuracy drop for Gemini 1.5 Flash over nine turns, while combining this prompt with a semantically equivalent reworded question caused a 7.5% drop for Claude 3.5 Haiku. Additionally, we find that model accuracy across turns can be effectively modeled using Markov chains, enabling the prediction of accuracy probabilities over time. This allows for estimation of the model's stationary (long-run) accuracy, which we find to be on average approximately 8% lower than its first-turn accuracy for Gemini 1.5 Flash. Our results from a model's hidden states also reveal evidence that linear probes can help predict future answer changes. Together, these results establish stationary accuracy as a principled robustness metric for interactive settings and expose the fragility of models under repeated questioning. Addressing this instability will be essential for deploying LLMs in high-stakes and interactive settings where consistent reasoning is as important as initial accuracy.

Shourya Batra, Pierce Tillman, Samarth Gaggar et al.

As Large Language Models (LLMs) evolve into personal assistants with access to sensitive user data, they face a critical privacy challenge: while prior work has addressed output-level privacy, recent findings reveal that LLMs often leak private information through their internal reasoning processes, violating contextual privacy expectations. These leaky thoughts occur when models inadvertently expose sensitive details in their reasoning traces, even when final outputs appear safe. The challenge lies in preventing such leakage without compromising the model's reasoning capabilities, requiring a delicate balance between privacy and utility. We introduce Steering Activations towards Leakage-free Thinking (SALT), a lightweight test-time intervention that mitigates privacy leakage in model's Chain of Thought (CoT) by injecting targeted steering vectors into hidden state. We identify the high-leakage layers responsible for this behavior. Through experiments across multiple LLMs, we demonstrate that SALT achieves reductions including $18.2\%$ reduction in CPL on QwQ-32B, $17.9\%$ reduction in CPL on Llama-3.1-8B, and $31.2\%$ reduction in CPL on Deepseek in contextual privacy leakage dataset AirGapAgent-R while maintaining comparable task performance and utility. Our work establishes SALT as a practical approach for test-time privacy protection in reasoning-capable language models, offering a path toward safer deployment of LLM-based personal agents.

Snigdha Pandya, Rohan Nagale, Kenji Sahay et al.

Large language models (LLMs) often generate fluent but factually incorrect statements despite having access to relevant evidence, a failure mode rooted in how they allocate attention between contextual and parametric knowledge. Understanding and steering this internal behavior is key both for trustworthy deployment and for scientific interpretability of model mechanisms. We introduce COMPASS (Context-Modulated PID Attention Steering System), a lightweight, interpretable control framework that embeds a model-based feedback loop directly within decoding. COMPASS quantifies context reliance via a transparent metric, the Context Reliance Score (CRS), which serves as an online probe of how attention heads ground generation in evidence. Using this interpretable signal, a PID controller dynamically modulates attention heads to maintain factual consistency without retraining or multi-pass decoding. Across benchmarks (HotpotQA, XSum, HaluEval, RAGTruth), COMPASS consistently reduces contextual hallucination rates (2.8 to 5.8 percent absolute) while revealing how distinct attention heads contribute to evidence alignment. These results highlight feedback-driven interpretability as a pathway toward scientific understanding of LLM behavior.

Sophie Li, Nicholas Huang, Nayan Saxena et al.

Large language models (LLMs) improve reasoning accuracy when generating multiple candidate solutions at test time, but standard methods like Best-of-N (BoN) incur high computational cost by fully generating all branches. Self-Truncation Best-of-N (ST-BoN) mitigates this by truncating unpromising paths early, but its reliance on consistency-based heuristics is a limitation as it does not directly evaluate branch quality. We present KL-Adjusted Pruned Path Algorithm (KAPPA), an inference-time method that combines Kullback-Leibler divergence, confidence, and entropy into a principled scoring function to guide progressive pruning. By promoting diversity during exploration and selectively eliminating low-scoring branches, KAPPA maintains accuracy while substantially reducing memory and token usage. Experiments on GSM8K and MATH500 with DeepSeek-R1-Distill-Qwen-1.5B and Qwen2.5-7B-Instruct demonstrate that KAPPA stabilizes performance in smaller models and achieves up to ~60% reduction in peak memory and ~90% reduction in total token generation relative to BoN, with minimal impact on accuracy.

Patricia Dao, Jashmitha Sappa, Saanvi Terala et al.

Traditional crime prediction techniques are slow and inefficient when generating predictions as crime increases rapidly \cite{r15}. To enhance traditional crime prediction methods, a Long Short-Term Memory and Gated Recurrent Unit model was constructed using datasets involving gender ratios, high school graduation rates, political status, unemployment rates, and median income by state over multiple years. While there may be other crime prediction tools, personalizing the model with hand picked factors allows a unique gap for the project. Producing an effective model would allow policymakers to strategically allocate specific resources and legislation in geographic areas that are impacted by crime, contributing to the criminal justice field of research \cite{r2A}. The model has an average total loss value of 70.792.30, and a average percent error of 9.74 percent, however both of these values are impacted by extreme outliers and with the correct optimization may be corrected.

Ishneet Sukhvinder Singh, Ritvik Aggarwal, Ibrahim Allahverdiyev et al.

Retrieval-Augmented Generation (RAG) systems using large language models (LLMs) often generate inaccurate responses due to the retrieval of irrelevant or loosely related information. Existing methods, which operate at the document level, fail to effectively filter out such content. We propose LLM-driven chunk filtering, ChunkRAG, a framework that enhances RAG systems by evaluating and filtering retrieved information at the chunk level. Our approach employs semantic chunking to divide documents into coherent sections and utilizes LLM-based relevance scoring to assess each chunk's alignment with the user's query. By filtering out less pertinent chunks before the generation phase, we significantly reduce hallucinations and improve factual accuracy. Experiments show that our method outperforms existing RAG models, achieving higher accuracy on tasks requiring precise information retrieval. This advancement enhances the reliability of RAG systems, making them particularly beneficial for applications like fact-checking and multi-hop reasoning.

Joshua Liu, Aarav Jain, Soham Takuri et al.

Rapid improvements in large language models have unveiled a critical challenge in human-AI interaction: sycophancy. In this context, sycophancy refers to the tendency of models to excessively agree with or flatter users, often at the expense of factual accuracy. While previous studies have primarily analyzed this behavior in single-turn interactions, its persistence and evolution in multi-step conversations remain largely unexplored. We introduce TRUTH DECAY, a benchmark specifically designed to evaluate sycophancy in extended dialogues, where language models must navigate iterative user feedback, challenges, and persuasion. We prompt models to elicit four types of sycophantic biases. We then propose and test sycophancy reduction strategies, evaluating their effectiveness beyond single-step interactions.

Joshua Lee, Wyatt Fong, Alexander Le et al.

Sarcasm detection is a significant challenge in sentiment analysis due to the nuanced and context-dependent nature of verbiage. We introduce Pragmatic Metacognitive Prompting (PMP) to improve the performance of Large Language Models (LLMs) in sarcasm detection, which leverages principles from pragmatics and reflection helping LLMs interpret implied meanings, consider contextual cues, and reflect on discrepancies to identify sarcasm. Using state-of-the-art LLMs such as LLaMA-3-8B, GPT-4o, and Claude 3.5 Sonnet, PMP achieves state-of-the-art performance on GPT-4o on MUStARD and SemEval2018. This study demonstrates that integrating pragmatic reasoning and metacognitive strategies into prompting significantly enhances LLMs' ability to detect sarcasm, offering a promising direction for future research in sentiment analysis.

David Puertolas Merenciano, Ekaterina Vasyagina, Raghav Dixit et al.

LoRA adapters let users fine-tune large language models (LLMs) efficiently. However, LoRA adapters are shared through open repositories like Hugging Face Hub \citep{huggingface_hub_docs}, making them vulnerable to backdoor attacks. Current detection methods require running the model with test input data -- making them impractical for screening thousands of adapters where the trigger for backdoor behavior is unknown. We detect poisoned adapters by analyzing their weight matrices directly, without running the model -- making our method data-agnostic. Our method extracts simple statistics -- how concentrated the singular values are, their entropy, and the distribution shape -- and flags adapters that deviate from normal patterns. We evaluate the method on 500 LoRA adapters -- 400 clean, and 100 poisoned for Llama-3.2-3B on instruction and reasoning datasets: Alpaca, Dolly, GSM8K, ARC-Challenge, SQuADv2, NaturalQuestions, HumanEval, and GLUE dataset. We achieve 97\% detection accuracy with less than 2\% false positives.

Logan Mann, Nayan Saxena, Sarah Tandon et al.

Negation instructions such as 'do not mention $X$' can paradoxically increase the accessibility of $X$ in human thought, a phenomenon known as ironic rebound. Large language models (LLMs) face the same challenge: suppressing a concept requires internally activating it, which may prime rebound instead of avoidance. We investigated this tension with two experiments. \textbf{(1) Load \& content}: after a negation instruction, we vary distractor text (semantic, syntactic, repetition) and measure rebound strength. \textbf{(2) Polarity separation}: We test whether models distinguish neutral from negative framings of the same concept and whether this separation predicts rebound persistence. Results show that rebound consistently arises immediately after negation and intensifies with longer or semantic distractors, while repetition supports suppression. Stronger polarity separation correlates with more persistent rebound. Together, these findings, complemented by a circuit tracing analysis that identifies sparse middle-layer attention heads amplifying forbidden tokens while early layers suppress, link cognitive predictions of ironic rebound with mechanistic insights into long-context interference. To support future work, we release ReboundBench, a dataset of $5,000$ systematically varied negation prompts designed to probe rebound in LLMs.

Dev Patel, Gabrielle Gervacio, Diekola Raimi et al.

Large Language Models require substantial computational resources for inference, posing deployment challenges. While dynamic pruning offers superior efficiency over static methods through adaptive circuit selection, it exacerbates alignment degradation by retaining only input-dependent safety-critical circuit preservation across diverse inputs. As a result, addressing these heightened alignment vulnerabilities remains critical. We introduce Alignment-Aware Probe Pruning (AAPP), a dynamic structured pruning method that adaptively preserves alignment-relevant circuits during inference, building upon Probe Pruning. Experiments on LLaMA 2-7B, Qwen2.5-14B-Instruct, and Gemma-3-12B-IT show AAPP improves refusal rates by 50\% at matched compute, enabling efficient yet safety-preserving LLM deployment.

Abhishek More, Anthony Zhang, Nicole Bonilla et al.

Chain-of-thought (CoT) prompting enables Large Language Models to solve complex problems, but deploying these models safely requires reliable confidence estimates, a capability where existing methods suffer from poor calibration and severe overconfidence on incorrect predictions. We propose Enhanced Dirichlet and Topology Risk (EDTR), a novel decoding strategy that combines topological analysis with Dirichlet-based uncertainty quantification to measure LLM confidence across multiple reasoning paths. EDTR treats each CoT as a vector in high-dimensional space and extracts eight topological risk features capturing the geometric structure of reasoning distributions: tighter, more coherent clusters indicate higher confidence while dispersed, inconsistent paths signal uncertainty. We evaluate EDTR against three state-of-the-art calibration methods across four diverse reasoning benchmarks spanning olympiad-level mathematics (AIME), grade school math (GSM8K), commonsense reasoning, and stock price prediction \cite{zhang2025aime, cobbe2021training, talmor-etal-2019-commonsenseqa, yahoo_finance}. EDTR achieves 41\% better calibration than competing methods with an average ECE of 0.287 and the best overall composite score of 0.672, while notably achieving perfect accuracy on AIME and exceptional calibration on GSM8K with an ECE of 0.107, domains where baselines exhibit severe overconfidence. Our work provides a geometric framework for understanding and quantifying uncertainty in multi-step LLM reasoning, enabling more reliable deployment where calibrated confidence estimates are essential.

Muna Numan Said, Aarib Zaidi, Rabia Usman et al.

The transformative potential of text-to-image (T2I) models hinges on their ability to synthesize culturally diverse, photorealistic images from textual prompts. However, these models often perpetuate cultural biases embedded within their training data, leading to systemic misrepresentations. This paper benchmarks the Component Inclusion Score (CIS), a metric designed to evaluate the fidelity of image generation across cultural contexts. Through extensive analysis involving 2,400 images, we quantify biases in terms of compositional fragility and contextual misalignment, revealing significant performance gaps between Western and non-Western cultural prompts. Our findings underscore the impact of data imbalance, attention entropy, and embedding superposition on model fairness. By benchmarking models like Stable Diffusion with CIS, we provide insights into architectural and data-centric interventions for enhancing cultural inclusivity in AI-generated imagery. This work advances the field by offering a comprehensive tool for diagnosing and mitigating biases in T2I generation, advocating for more equitable AI systems.