Search: “safety”

Cross-Modal Attention Analysis and Optimization in Vision-Language Models: A Study on Visual Reliability

arXiv:2604.17217v1 Announce Type: new Abstract: Vision-Language Models (VLMs) achieve strong cross-modal performance, yet recent evidence suggests they over-rely on textual descriptions while under-ut

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

CrossFlowDG: Bridging the Modality Gap with Cross-modal Flow Matching for Domain Generalization

DGX agent

arXiv:2604.16892v1 Announce Type: new Abstract: Domain generalization (DG) aims to maintain performance under domain shift, which in computer vision appears primarily as stylistic variations that caus

safetyarxiv-cs-cv

21 Apr 2026

Culture-Aware Humorous Captioning: Multimodal Humor Generation across Cultural Contexts

arXiv:2604.18091v1 Announce Type: new Abstract: Recent multimodal large language models have shown promising ability in generating humorous captions for images, yet they still lack stable control over

Model Releases

DART: Mitigating Harm Drift in Difference-Aware LLMs via Distill-Audit-Repair Training

model-releasesarxiv-cs-cl

arXiv:2604.16845v1 Announce Type: new Abstract: Large language models (LLMs) tuned for safety often avoid acknowledging demographic differences, even when such acknowledgment is factually correct (e.g

Debate as Reward: A Multi-Agent Reward System for Scientific Ideation via RL Post-Training

arXiv:2604.16723v1 Announce Type: cross Abstract: Large Language Models (LLMs) have demonstrated potential in automating scientific ideation, yet current approaches relying on iterative prompting or c

Decoding AI Tutor Effects for Educational Measurement: Temporal, Multi-Outcome, and Behavior-Cognitive Analysis

arXiv:2604.16366v1 Announce Type: cross Abstract: Artificial intelligence (AI) tutors have become increasingly popular in learning environments. In this study, we propose an AI agent prototype framewo

Decoupling the Effect of Chain-of-Thought Reasoning: A Human Label Variation Perspective

arXiv:2601.03154v2 Announce Type: replace Abstract: Reasoning-tuned LLMs utilizing long Chain-of-Thought (CoT) excel at single-answer tasks, yet their ability to model Human Label Variation--which req

DeepThinkVLA: Enhancing Reasoning Capability of Vision-Language-Action Models

arXiv:2511.15669v2 Announce Type: replace Abstract: Does Chain-of-Thought (CoT) reasoning genuinely improve Vision-Language-Action (VLA) models, or does it merely add overhead? Existing CoT-VLA system

Demystifying the unreasonable effectiveness of online alignment methods

arXiv:2604.17207v1 Announce Type: cross Abstract: Iterative alignment methods based on purely greedy updates are remarkably effective in practice, yet existing theoretical guarantees of (O(log T)) KL-

DFedReweighting: A Unified Framework for Objective-Oriented Reweighting in Decentralized Federated Learning

arXiv:2512.12022v2 Announce Type: replace Abstract: Decentralized federated learning (DFL) has emerged as a promising paradigm that enables multiple clients to collaboratively train machine learning m

Diagnosing LLM-based Rerankers in Cold-Start Recommender Systems: Coverage, Exposure and Practical Mitigations

arXiv:2604.16318v1 Announce Type: cross Abstract: Large language models (LLMs) and cross-encoder rerankers have gained attention for improving recommender systems, particularly in cold-start scenarios

DiffCoT: Diffusion-styled Chain-of-Thought Reasoning in LLMs

arXiv:2601.03559v2 Announce Type: replace Abstract: Chain-of-Thought (CoT) reasoning improves multi-step mathematical problem solving in large language models but remains vulnerable to exposure bias a

Differential Privacy in Two-Layer Networks: How DP-SGD Harms Fairness and Robustness

arXiv:2603.04881v2 Announce Type: replace Abstract: Differentially private learning is essential for training models on sensitive data, but empirical studies consistently show that it can degrade perf

Distributional Off-Policy Evaluation with Deep Quantile Process Regression

arXiv:2604.18143v1 Announce Type: cross Abstract: This paper investigates the off-policy evaluation (OPE) problem from a distributional perspective. Rather than focusing solely on the expectation of t

Diverse Dictionary Learning

arXiv:2604.17568v1 Announce Type: new Abstract: Given only observational data X = g(Z), where both the latent variables Z and the generating process g are unknown, recovering Z is ill-posed without ad

DMax: Aggressive Parallel Decoding for dLLMs

arXiv:2604.08302v2 Announce Type: replace Abstract: We present DMax, a new paradigm for efficient diffusion language models (dLLMs). It mitigates error accumulation in parallel decoding, enabling aggr

Do LLMs Use Cultural Knowledge Without Being Told? A Multilingual Evaluation of Implicit Pragmatic Adaptation

arXiv:2604.17718v1 Announce Type: new Abstract: Many benchmarks show that large language models can answer direct questions about culture. We study a different question: do they also change how they s

Does 'Do Differentiable Simulators Give Better Policy Gradients?'' Give Better Policy Gradients?

arXiv:2604.18161v1 Announce Type: new Abstract: In policy gradient reinforcement learning, access to a differentiable model enables 1st-order gradient estimation that accelerates learning compared to

Does Welsh media need a review? Detecting bias in Nation.Cymru's political reporting

arXiv:2604.17628v1 Announce Type: new Abstract: Wales' political landscape has been marked by growing accusations of bias in Welsh media. This paper takes the first computational step toward testing t

DOSE: Data Selection for Multi-Modal LLMs via Off-the-Shelf Models

arXiv:2604.16979v1 Announce Type: cross Abstract: High-quality and diverse multimodal data are essential for improving vision-language models (VLMs), yet existing datasets often contain noisy, redunda

DR-SAC: Distributionally Robust Soft Actor-Critic for Reinforcement Learning under Uncertainty

arXiv:2506.12622v2 Announce Type: replace Abstract: Deep reinforcement learning (RL) has achieved remarkable success, yet its deployment in real-world scenarios is often limited by vulnerability to en

DreamShot: Personalized Storyboard Synthesis with Video Diffusion Prior

arXiv:2604.17195v1 Announce Type: new Abstract: Storyboard synthesis plays a crucial role in visual storytelling, aiming to generate coherent shot sequences that visually narrate cinematic events with

Dynamic Emotion and Personality Profiling for Multimodal Deception Detection

arXiv:2604.17037v1 Announce Type: new Abstract: Deception detection is of great significance for ensuring information security and conducting public opinion analysis, with personality factors and emot

Dynamic Visual-semantic Alignment for Zero-shot Learning with Ambiguous Labels

arXiv:2604.17710v1 Announce Type: new Abstract: Zero-shot learning (ZSL) aims to recognize unseen classes without visual instances. However, existing methods usually assume clean labels, overlooking r

DynaWeb: Model-Based Reinforcement Learning of Web Agents

arXiv:2601.22149v2 Announce Type: replace Abstract: The development of autonomous web agents, powered by Large Language Models (LLMs) and reinforcement learning (RL), represents a significant step tow

Efficient Federated RLHF via Zeroth-Order Policy Optimization

arXiv:2604.17747v1 Announce Type: new Abstract: This paper considers reinforcement learning from human feedback in a federated learning setting with resource-constrained agents, such as edge devices.

Empowering Multi-Turn Tool-Integrated Agentic Reasoning with Group Turn Policy Optimization

arXiv:2511.14846v2 Announce Type: replace-cross Abstract: Training Large Language Models (LLMs) for multi-turn Tool-Integrated Reasoning (TIR) - where models iteratively reason, generate code, and ver

End-to-End Optimization of LLM-Driven Multi-Agent Search Systems via Heterogeneous-Group-Based Reinforcement Learning

arXiv:2506.02718v2 Announce Type: replace Abstract: Large language models (LLMs) are versatile, yet their deployment in complex real-world settings is limited by static knowledge cutoffs and the diffi

EVE: Verifiable Self-Evolution of MLLMs via Executable Visual Transformations

arXiv:2604.18320v1 Announce Type: new Abstract: Self-evolution of multimodal large language models (MLLMs) remains a critical challenge: pseudo-label-based methods suffer from progressive quality degr

Evidence-Augmented Policy Optimization with Reward Co-Evolution for Long-Context Reasoning

arXiv:2601.10306v2 Announce Type: replace-cross Abstract: While Reinforcement Learning (RL) has advanced LLM reasoning, applying it to long-context scenarios is hindered by sparsity of outcome rewards

Evolutionary Negative Module Pruning for Better LoRA Merging

arXiv:2604.17753v1 Announce Type: cross Abstract: Merging multiple Low-Rank Adaptation (LoRA) experts into a single backbone is a promising approach for efficient multi-task deployment. While existing

Explanation Bias is a Product: Revealing the Hidden Lexical and Position Preferences in Post-Hoc Feature Attribution

arXiv:2512.11108v3 Announce Type: replace Abstract: Good quality explanations strengthen the understanding of language models and data. Feature attribution methods, such as Integrated Gradient, are a

FairLogue: Evaluating Intersectional Fairness across Clinical Machine Learning Use Cases using the All of Us Research Program

arXiv:2604.16450v1 Announce Type: cross Abstract: Intersectional biases in healthcare data can produce compound disparities in clinical machine learning models, yet most fairness evaluations assess de

Fairness Constraints in High-Dimensional Generalized Linear Models

arXiv:2604.16610v1 Announce Type: cross Abstract: Machine learning models often inherit biases from historical data, raising critical concerns about fairness and accountability. Conventional fairness

FairNVT: Improving Fairness via Noise Injection in Vision Transformers

arXiv:2604.16780v1 Announce Type: new Abstract: This paper presents FairNVT, a lightweight debiasing framework for pretrained transformer-based encoders that improves both representation and predictio

'Faithful to What?' On the Limits of Fidelity-Based Explanations

arXiv:2506.12176v5 Announce Type: replace Abstract: In explainable AI, surrogate models are commonly evaluated by their fidelity to a neural network's predictions. Fidelity, however, measures alignmen

Fisher Decorator: Refining Flow Policy via A Local Transport Map

arXiv:2604.17919v1 Announce Type: new Abstract: Recent advances in flow-based offline reinforcement learning (RL) have achieved strong performance by parameterizing policies via flow matching. However

Foundation Model for Cardiac Time Series via Masked Latent Attention

arXiv:2603.26475v2 Announce Type: replace Abstract: Electrocardiograms (ECGs) are among the most widely available clinical signals and play a central role in cardiovascular diagnosis. While recent fou

Function Words as Statistical Cues for Language Learning

arXiv:2601.21191v2 Announce Type: replace Abstract: What statistical properties might support learning abstract grammatical knowledge from linear input? We address this question by examining the stati

Generative Semantic Communication via Alternating Dual-Domain Posterior Sampling

arXiv:2604.16796v1 Announce Type: new Abstract: Generative semantic communication (SemCom) harnesses pretrained generative priors to improve the perceptual quality of wireless image transmission. Exis

Geometric Stability: The Missing Axis of Representations

arXiv:2601.09173v4 Announce Type: replace-cross Abstract: Representational similarity analysis and related methods have become standard tools for comparing the internal geometries of neural networks a

GeometryZero: Advancing Geometry Solving via Group Contrastive Policy Optimization

arXiv:2506.07160v3 Announce Type: replace Abstract: Recent progress in large language models (LLMs) has boosted mathematical reasoning, yet geometry remains challenging where auxiliary construction is

GS-STVSR: Ultra-Efficient Continuous Spatio-Temporal Video Super-Resolution via 2D Gaussian Splatting

arXiv:2604.18047v1 Announce Type: new Abstract: Continuous Spatio-Temporal Video Super-Resolution (C-STVSR) aims to simultaneously enhance the spatial resolution and frame rate of videos by arbitrary

HEALing Entropy Collapse: Enhancing Exploration in Few-Shot RLVR via Hybrid-Domain Entropy Dynamics Alignment

arXiv:2604.17928v1 Announce Type: new Abstract: Reinforcement Learning with Verifiable Reward (RLVR) has proven effective for training reasoning-oriented large language models, but existing methods la

High-Resolution Visual Reasoning via Multi-Turn Grounding-Based Reinforcement Learning

arXiv:2507.05920v2 Announce Type: replace Abstract: State-of-the-art large multi-modal models (LMMs) face challenges when processing high-resolution images, as these inputs are converted into enormous

How Language Models Conflate Logical Validity with Plausibility: A Representational Analysis of Content Effects

arXiv:2510.06700v3 Announce Type: replace Abstract: Both humans and large language models (LLMs) exhibit content effects: biases in which the plausibility of the semantic content of a reasoning proble

Human-Centered Supervision for Sentiment Analysis in Telugu: A Systematic Inquiry Beyond Accuracy

arXiv:2508.01486v3 Announce Type: replace Abstract: Sentiment analysis for low-resource languages remains challenging in an era where interpretability, human alignment, and fairness are increasingly n

Hybrid Multi-Dimensional MRI Prostate Cancer Detection via Hadamard Network-Based Bias Correction and Residual Networks

arXiv:2604.17107v1 Announce Type: new Abstract: Magnetic Resonance Imaging (MRI) is vital for prostate cancer (PCa) diagnosis. While advanced techniques such as Hybrid Multi-dimensional MRI (HM-MRI) h