Search arXiv⌕ Search

arXiv subjects

Chirag Agarwal

Publications and source records attributed to Chirag Agarwal.

At least 19 recordsLinked to original sources
arXiv2026

The Fragility of Chain-of-Thought Monitoring Across Typologically Diverse Languages

Chain-of-thought (CoT) monitoring has been proposed as a promising safety mechanism for detecting misaligned behavior in large language models. However, its reliability remains largely unexplored beyond English and across diverse model families. We present the first large-scale evaluation of CoT monitorability across 13 diverse languages and seven frontier model families, comprising 16 models. Using adversarial-hint evaluations that require explicit intermediate computation, together with analysis of internal answer-token probabilities, we consistently find CoT unfaithfulness across languages and hint types, with an average rate of 95.9\% across 8B--120B parameter models. We find that frontier models systematically exhibit strategic manipulation, including answer-switching, post-hoc rationalization, and procedural exploitation of hints, making their reasoning difficult to reliably monitor. These deceptive patterns remain especially pronounced in low-resource languages, revealing fundamental limitations in current CoT-based oversight. Our results show that CoT monitoring is fragile under linguistic distribution shift, providing a substantially weaker safety signal than English-only studies suggest. These findings motivate the development of more robust CoT monitors and complementary white-box monitoring techniques, particularly for mid- and low-resource languages. Our code is available \href{https://multilingual-cot-monitoring.github.io/}{\textcolor{blue}{here}}.

cs.CL↗
arXiv2026

STRUCTUREDAGENT: Planning with AND/OR Trees for Long-Horizon Web Tasks

Existing LLM-based web agents struggle on complex, long-horizon tasks due to limited in-context memory, weak planning abilities, and greedy behaviors that lead to premature termination. To address these challenges, we propose \SA{}, a hierarchical planning framework that interleaves planning and execution via dynamic $\ANDOR$ trees. The framework separates structural planning from LLM-based reasoning, enabling principled error recovery through node repair, systematic exploration of alternatives via OR nodes, and modular plans that can facilitate human intervention. On WebArena (630 tasks), \SA{} achieves a $\sim$53\% success rate vs.\ $\sim$46\% for AgentOccam, and on complex multi-constraint Amazon shopping tasks, gains reach 10\% over the strongest baseline.

cs.AI↗
arXiv2026

PageGuide: Browser extension to assist users in navigating a webpage and locating information

Users browsing the web struggle to locate relevant information on cluttered pages and to complete web navigation tasks. Current web agents can answer questions and automate actions, but return answers without showing where the information comes from, forcing users to manually verify results and blindly trust every automated step. We present PageGuide, a browser extension that grounds LLM answers in page content (e.g. HTML DOM elements, text, maps, images, charts), via two features: FIND, which highlights supporting evidence in situ on a page, and GUIDE, which guides users across pages with evidence captured at each step. In a within-subject study (N=92), PageGuide improves overall verification accuracy from 65% to 77% and reduces verification time from 73s to 67s across FIND and GUIDE tasks. PageGuide also reduces manual search effort and improves users' perceived support for verification. Code is available at https://pageguide.github.io/.

cs.HC↗
arXiv2026

The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing

Autonomous AI systems are transitioning from advisory roles to autonomous ones for medication prescriptions. Recent U.S. bill H.R. 238 and Utah's prescription-renewal pilot program both authorize AI to prescribe medications in an agentic capacity. While many regulatory guidelines suggest aggregate model performance metrics at the point of clearance, they do not require i) calibrated per-prediction confidence for action-gated thresholds, ii) differentiated communication between uncertainty arising from model ignorance (epistemic) from genuine clinical ambiguity (aleatoric), and iii) inferential transparency at the moment of decision enabling liability allocation. Here, we argue these three architectural features are minimum conditions for safe autonomous prescribing, and validate them with a survey of 136 U.S. prescribing clinicians. Our results suggest prescribing clinicians i) would not permit autonomous prescribing without a confidence-based escalation mechanism, ii) preferred a competing-options summary for aleatoric uncertainty but preferred abstention for epistemic uncertainty, and iii) were only willing to accept liability when inferential transparency enabled them to make a decision under acknowledged uncertainty. These findings indicate that our recommended architectural features would encourage higher rates of clinician adoption of autonomous AI prescribing, largely through collapsing much of what "autonomy" conventionally means.

cs.AI↗
arXiv2026

A Mechanistic Perspective and Circuit-Guided Difficulty Metric for Unlearning

Machine unlearning is becoming essential for building trustworthy and compliant language models. Yet unlearning success varies considerably across individual samples: some are reliably erased, while others persist despite the same procedure. We argue that this disparity is not only a data-side phenomenon, but also reflects model-internal mechanisms that encode and protect memorized information. We study this problem from a mechanistic perspective based on model circuits--structured interaction pathways that govern how predictions are formed. We propose Circuit-guided Unlearning Difficulty (CUD), a {\em pre-unlearning} metric that assigns each sample a continuous difficulty score using circuit-level signals. Extensive experiments demonstrate that CUD reliably separates intrinsically easy and hard samples, and remains stable across unlearning methods. We identify key circuit-level patterns that reveal a mechanistic signature of difficulty: easy-to-unlearn samples are associated with shorter, shallower interactions concentrated in earlier-to-intermediate parts of the original model, whereas hard samples rely on longer and deeper pathways closer to late-stage computation. Compared to existing qualitative studies, CUD takes a first step toward a principled, fine-grained, and interpretable analysis of unlearning difficulty; and motivates the development of unlearning methods grounded in model mechanisms.

cs.LG↗
arXiv2026

When Graph Tokens Sink: A Mechanistic Analysis of Graph Language Models

Graph Language Models (GLMs) have become a promising direction for adapting Large Language Models (LLMs) to graph learning tasks. By transforming graph topology and node information into graph tokens, GLMs allow LLMs to jointly process structured graph inputs and textual instructions. Yet, it remains unclear how LLMs internally interpret these graph tokens and whether graph tokens act as meaningful carriers of graph structure. In this work, we analyze how LLMs process graph information through graph-token behavior in representative GLM architectures. Findings. We find that the internal saliency of graph tokens in GLMs is not equivalent to graph information utilization. Graph sink tokens consistently emerge as activation-level outliers: they can be identified by massive activation values along a small set of hidden-state dimensions and are biased toward early graph-token positions. However, this activation-level saliency does not imply that these tokens are the main carriers of graph information. Unlike classical attention sinks in language and vision-language models, graph sink tokens do not necessarily attract the largest attention weights from query tokens. Through pruning, repositioning, and swapping interventions, we show that graph sink tokens are not the most important semantic or structural tokens for downstream prediction. Implications. Together, these results suggest that after current GLMs map graph structure into the LLM token space, the resulting graph-token representations do not naturally form a fully usable topology-aware internal representation; instead, they exhibit a decoupling between activation-level saliency and graph-semantic utility. This decoupling points to limitations in existing graph-token construction, placement, and alignment mechanisms.

cs.LG↗
arXiv2026

GridVQA-X: A Framework for Evaluating Multimodal Explainability Methods

With the increasing development of Vision-Language Models, it becomes imperative that their predictions are readily explainable to relevant stakeholders. However, the field of explainability has not kept pace with the multimodal surge. While recent Multimodal Explainable AI (MxAI) methods generate explanations to attribute the interaction between different modalities, current evaluation protocols lack the ground truth required to distinguish between true cross-modal reasoning (e.g., spatial composition) and shallow cross-modal shortcuts (e.g., Bag-of-Words attribute matching). It remains unknown whether MxAI methods faithfully capture synergistic interactions or merely hallucinate reasoning on models acting as simple feature detectors. In this paper, we introduce GridVQA-X, the first diagnostic framework specifically designed to evaluate cross-modal explainability. Unlike natural datasets, GridVQA-X leverages a closed-world synthesis logic to generate unique, mathematically guaranteed explanations. We utilize this controlled environment to train paired ground-truth models on identical architectures: $M_{\text{pure}}$, which learns robust spatial-relational reasoning and $M_{\text{spur}}$, which is structurally forced to rely on cross-modal shortcuts. This behavioral divergence creates a rigorous testbed: a faithful explainer must report distinct reasoning pathways for each model. Our findings reveal that widely used methods fail to distinguish between models relying on genuine spatial-relational reasoning and those exploiting cross-modal shortcuts, highlighting a critical gap in capturing true cross-modal synergy and misrepresenting how multimodal models actually make decisions.

cs.CV↗
arXiv2026

CURE-Med: Curriculum-Informed Reinforcement Learning for Multilingual Medical Reasoning

While large language models (LLMs) have shown to perform well on monolingual mathematical and commonsense reasoning, they remain unreliable for multilingual medical reasoning applications, hindering their deployment in multilingual healthcare settings. We address this by first introducing CUREMED-BENCH, a high-quality multilingual medical reasoning dataset with open-ended reasoning queries with a single verifiable answer, spanning thirteen languages, including underrepresented languages such as Amharic, Yoruba, and Swahili. Building on this dataset, we propose CURE-MED, a curriculum-informed reinforcement learning framework that integrates code-switching-aware supervised fine-tuning and Group Relative Policy Optimization to jointly improve logical correctness and language stability. Across thirteen languages, our approach consistently outperforms strong baselines and scales effectively, achieving 85.21% language consistency and 54.35% logical correctness at 7B parameters, and 94.96% language consistency and 70.04% logical correctness at 32B parameters. These results support reliable and equitable multilingual medical reasoning in LLMs. The code and dataset are available at https://cure-med.github.io/

cs.AI↗
arXiv2026

Towards Understanding the Robustness of Sparse Autoencoders

Large Language Models (LLMs) remain vulnerable to optimization-based jailbreak attacks that exploit internal gradient structure. While Sparse Autoencoders (SAEs) are widely used for interpretability, their robustness implications remain underexplored. We present a study of integrating pretrained SAEs into transformer residual streams at inference time, without modifying model weights or blocking gradients. Across four model families (Gemma, LLaMA, Mistral, Qwen) and two strong white-box attacks (GCG, BEAST) plus three black-box benchmarks, SAE-augmented models achieve up to a 5x reduction in jailbreak success rate relative to the undefended baseline and reduce cross-model attack transferability. Parametric ablations reveal (i) a monotonic dose-response relationship between L0 sparsity and attack success rate, and (ii) a layer-dependent defense-utility tradeoff, where intermediate layers balance robustness and clean performance. These findings are consistent with a representational bottleneck hypothesis: sparse projection reshapes the optimization geometry exploited by jailbreak attacks.

cs.LG↗
arXiv2026

Do Vision Language Models Need to Process Image Tokens?

Vision Language Models (VLMs) have achieved remarkable success by integrating visual encoders with large language models (LLMs). While VLMs process dense image tokens across deep transformer stacks (incurring substantial computational overhead), it remains fundamentally unclear whether sustained image-token processing is necessary for their performance or visual representations meaningfully evolve from early to later layers. In this work, we systematically investigate the functional role of image tokens in VLMs and show that visual representations rapidly converge to a bounded-complexity regime, \ie their entropy stabilizes, intrinsic dimensionality compresses, and trajectory curvature approaches a near-constant profile. In contrast, textual representations continue to undergo substantial restructuring across depth. Once stabilized, visual representations become largely interchangeable between layers, indicating limited additional transformation in deeper stages. Further, depth-wise visual truncation reveals that the necessity of visual processing is task-dependent, where single-token predictions remain comparatively robust to truncated visual depth, but multi-token generation require sustained access to visual representations. Under deterministic decoding, reducing visual depth perturbs intermediate reasoning trajectories more strongly than final outputs, suggesting that image tokens influence the structure of reasoning more than the ultimate conclusions. Collectively, these findings \textbf{question the assumption} that deeper visual processing is uniformly essential in VLMs, challenging the current paradigm of multimodal LLM architectures.

cs.CV↗
arXiv2026

SynopticBench: Evaluating Vision-Language Models on Generating Weather Forecast Discussions of the Future

Recent advances in visual-language models (VLMs) have led to significant improvements in a plethora of complex multimodal tasks like image captioning, report generation, and visual perception. However, generating text from meteorological data is highly challenging because the atmosphere is a chaotic system that is rapidly changing at various spatial and temporal scales. Given the complexity of atmospheric phenomena, it is critical to verifiably quantify the effectiveness of existing VLMs on weather forecasting data. In this work, we present SynopticBench, a high-quality dataset consisting of 1,367,041 text samples of Area Forecast Discussions created by the National Weather Service over the continental United States paired to images of 500mb geopotential height, 2 meter temperature, and 850mb wind velocity in weather forecasts. We also present Synoptic Phenomena Alignment and Coverage Evaluation (SPACE), a novel evaluation framework that can be used to effectively estimate the quality of text descriptions of synoptic weather phenomena. Extensive experiments on generating forecast discussions using state-of-the-art VLMs show the sensitivity of existing evaluation metrics in this domain and enable further exploration into synoptic weather and climate text generation.

cs.CL↗
arXiv2026

Quantifying Explanation Quality in Graph Neural Networks using Out-of-Distribution Generalization

Evaluating the quality of post-hoc explanations for Graph Neural Networks (GNNs) remains a significant challenge. While recent years have seen an increasing development of explainability methods, current evaluation metrics (e.g., fidelity, sparsity) often fail to assess whether an explanation identifies the true underlying causal variables. To address this, we propose the Explanation-Generalization Score (EGS), a metric that quantifies the causal relevance of GNN explanations. EGS is founded on the principle of feature invariance and posits that if an explanation captures true causal drivers, it should lead to stable predictions across distribution shifts. To quantify this, we introduce a framework that trains GNNs using explanatory subgraphs and evaluates their performance in Out-of-Distribution (OOD) settings (here, OOD generalization serves as a rigorous proxy for the explanation's causal validity). Through large-scale validation involving 11,200 model combinations across synthetic and real-world datasets, our results demonstrate that EGS provides a principled benchmark for ranking explainers based on their ability to capture causal substructures, offering a robust alternative to traditional fidelity-based metrics.

cs.LG↗
arXiv2026

Improving Human Verification of LLM Reasoning through Interactive Explanation Interfaces

The reasoning capabilities of Large Language Models (LLMs) have led to their increasing employment in several critical applications, particularly education, where they support problem-solving, tutoring, and personalized study. Chain-of-thought (CoT) reasoning capabilities [1, 2] are well-known to help LLMs decompose a problem into steps and explore the solution spaces more effectively, leading to impressive performance on mathematical and reasoning benchmarks. As the length of CoT tokens per question increases substantially to even thousands of tokens per question [ 1], it is unknown how users could comprehend LLM reasoning and detect errors or hallucinations. To address this problem and understand how reasoning can improve human-AI interaction, we present three new interactive reasoning interfaces: interactive CoT (iCoT), interactive Program-of-Thought (iPoT), and interactive Graph (iGraph). That is, we ask LLMs themselves to generate an interactive web interface wrapped around the original CoT content, which may be presented in text (iCoT), graphs (iGraph) or code (iPoT). This interface allows users to interact with and provide a novel experience in reading and validating the reasoning chains of LLMs. Across a study of 125 participants, interactive interfaces significantly improve user performance. Specifically, iGraph users score the highest error detection rate (85.6%), followed by iPoT (82.5%), iCoT (80.6%), all outperforming standard CoT (73.5%). Interactive interfaces also lead to faster user validation time-iGraph users are faster (57.9 secs per question) than the users of iCoT and iPoT (60 secs) and the standard CoT (64.7 secs). A post-study questionnaire shows that users prefer iGraph, citing its superior ability to enable them to follow the LLM's reasoning. We discuss the implications of these results and provide recommendations for the future design of reasoning models.

cs.HC↗
arXiv2025

CLINIC: Evaluating Multilingual Trustworthiness in Language Models for Healthcare

Integrating language models (LMs) in healthcare systems holds great promise for improving medical workflows and decision-making. However, a critical barrier to their real-world adoption is the lack of reliable evaluation of their trustworthiness, especially in multilingual healthcare settings. Existing LMs are predominantly trained in high-resource languages, making them ill-equipped to handle the complexity and diversity of healthcare queries in mid- and low-resource languages, posing significant challenges for deploying them in global healthcare contexts where linguistic diversity is key. In this work, we present CLINIC, a Comprehensive Multilingual Benchmark to evaluate the trustworthiness of language models in healthcare. CLINIC systematically benchmarks LMs across five key dimensions of trustworthiness: truthfulness, fairness, safety, robustness, and privacy, operationalized through 18 diverse tasks, spanning 15 languages (covering all the major continents), and encompassing a wide array of critical healthcare topics like disease conditions, preventive actions, diagnostic tests, treatments, surgeries, and medications. Our extensive evaluation reveals that LMs struggle with factual correctness, demonstrate bias across demographic and linguistic groups, and are susceptible to privacy breaches and adversarial attacks. By highlighting these shortcomings, CLINIC lays the foundation for enhancing the global reach and safety of LMs in healthcare across diverse languages.

cs.CL↗
arXiv2025

Polarity-Aware Probing for Quantifying Latent Alignment in Language Models

Advances in unsupervised probes such as Contrast-Consistent Search (CCS), which reveal latent beliefs without relying on token outputs, raise the question of whether these methods can reliably assess model alignment. We investigate this by examining the sensitivity of CCS to harmful vs. safe statements and by introducing Polarity-Aware CCS (PA-CCS), a method for evaluating whether a model's internal representations remain consistent under polarity inversion. We propose two alignment-oriented metrics, Polar-Consistency and the Contradiction Index, to quantify the semantic robustness of a model's latent knowledge. To validate PA-CCS, we curate two main datasets and one control dataset containing matched harmful-safe sentence pairs constructed using different methodologies (concurrent and antagonistic statements). We apply PA-CCS to 16 language models. Our results show that PA-CCS identifies both architectural and layer-specific differences in the encoding of latent harmful knowledge. Notably, replacing the negation token with a meaningless marker degrades PA-CCS scores for models with well-aligned internal representations, while models lacking robust internal calibration do not exhibit this degradation. Our findings highlight the potential of unsupervised probing for alignment evaluation and emphasize the need to incorporate structural robustness checks into interpretability benchmarks. Code and datasets are available at: https://github.com/SadSabrina/polarity-probing. WARNING: This paper contains potentially sensitive, harmful, and offensive content.

cs.CL↗
arXiv2025

Do Students Debias Like Teachers? On the Distillability of Bias Mitigation Methods

Knowledge distillation (KD) is an effective method for model compression and transferring knowledge between models. However, its effect on model's robustness against spurious correlations that degrade performance on out-of-distribution data remains underexplored. This study investigates the effect of knowledge distillation on the transferability of ``debiasing'' capabilities from teacher models to student models on natural language inference (NLI) and image classification tasks. Through extensive experiments, we illustrate several key findings: (i) overall the debiasing capability of a model is undermined post-KD; (ii) training a debiased model does not benefit from injecting teacher knowledge; (iii) although the overall robustness of a model may remain stable post-distillation, significant variations can occur across different types of biases; and (iv) we pin-point the internal attention pattern and circuit that causes the distinct behavior post-KD. Given the above findings, we propose three effective solutions to improve the distillability of debiasing methods: developing high quality data for augmentation, implementing iterative knowledge distillation, and initializing student models with weights obtained from teacher models. To the best of our knowledge, this is the first study on the effect of KD on debiasing and its interenal mechanism at scale. Our findings provide understandings on how KD works and how to design better debiasing methods.

cs.LG↗
arXiv2025

A Survey of Multilingual Reasoning in Language Models

While reasoning and multilingual capabilities in language models (LMs) have achieved remarkable progress in recent years, their integration into a unified paradigm - multilingual reasoning - is at a nascent stage. Multilingual reasoning requires language models to handle logical reasoning across languages while addressing misalignment, biases, and challenges in low-resource settings. This survey provides the first in-depth review of multilingual reasoning in LMs. In this survey, we provide a systematic overview of existing methods that leverage LMs for multilingual reasoning, specifically outlining the challenges, motivations, and foundational aspects of applying language models to reason across diverse languages. We provide an overview of the standard data resources used for training multilingual reasoning in LMs and the evaluation benchmarks employed to assess their multilingual capabilities. Next, we analyze various state-of-the-art methods and their performance on these benchmarks. Finally, we explore future research opportunities to improve multilingual reasoning in LMs, focusing on enhancing their ability to handle diverse languages and complex reasoning tasks. Rapid growth of evolving developments in this field can be actively tracked on our project page: [https://github.com/AkashGhosh/Survey-of-Multilingual-Reasoning-in-Language-Models](https://github.com/AkashGhosh/Survey-of-Multilingual-Reasoning-in-Language-Models)

cs.CL↗
arXiv2025

Interpretable Neuropsychiatric Diagnosis via Concept-Guided Graph Neural Networks

Nearly one in five adolescents currently live with a diagnosed mental or behavioral health condition, such as anxiety, depression, or conduct disorder, underscoring the urgency of developing accurate and interpretable diagnostic tools. Resting-state functional magnetic resonance imaging (rs-fMRI) provides a powerful lens into large-scale functional connectivity, where brain regions are modeled as nodes and inter-regional synchrony as edges, offering clinically relevant biomarkers for psychiatric disorders. While prior works use graph neural network (GNN) approaches for disorder prediction, they remain complex black-boxes, limiting their reliability and clinical translation. In this work, we propose CONCEPTNEURO, a concept-based diagnosis framework that leverages large language models (LLMs) and neurobiological domain knowledge to automatically generate, filter, and encode interpretable functional connectivity concepts. Each concept is represented as a structured subgraph linking specific brain regions, which are then passed through a concept classifier. Our design ensures predictions through clinically meaningful connectivity patterns, enabling both interpretability and strong predictive performance. Extensive experiments across multiple psychiatric disorder datasets demonstrate that CONCEPTNEURO-augmented GNNs consistently outperform their vanilla counterparts, improving accuracy while providing transparent, clinically aligned explanations. Furthermore, concept analyses highlight disorder-specific connectivity patterns that align with expert knowledge and suggest new hypotheses for future investigation, establishing CONCEPTNEURO as an interpretable, domain-informed framework for psychiatric disorder diagnosis.

cs.LG↗