Research

Top Latest Rising Hidden gems

Today This week This month

Open-source gems →Radar →

Hugging Face Daily Papers4d ago

Explaining the Roles of Experts in Mixture-of-Experts LLMs for Personalization

DICL-Qwen3VL: Decoupled Confidence Calibration for Interpretable Workwear Detection

TransNSNet: A Transformer-Enhanced Multi-Scale Adversarial Framework for CT-to-PET Image Synthesis

Logic-Guided Graph Denoising Diffusion Models

Anisotropic Representation Dynamics in Transformers

Data-Efficient Classifier Pretraining Through Diffusion Models

Hugging Face Daily Papers4d ago

Behavioral Compositionality of Sparse Autoencoder Representations

Beyond SAEs: Consistent Concept Discovery with Sparse Coding

Learning to Be Guided: Shaping Latent Geometry for Classifier-Free Guidance

When Does Uncertainty Measure RAG Hallucination Risk?

Decomposing Brain-LLM Alignment with Interpretable Semantic Representations

SemPurify: Semantics-Aware Data Purification for Black-box Backdoor Defense

Spurious Attention Heads in Vision Transformers: Mechanistic Identification of Annotation-Driven Bias in Medical Chest X-Ray Analysis

SETAREH Transform: A Sorting-Assisted Haar-Wavelet Sparsification Method for Approximate Inner-Product Computation

CLEAR: Cooperative Latent Flow for Visual Counterfactual Explanations with Accurate Region Preservation

Escaping Parameter Space: Tight Generalization Bounds via Representation Quality

FedRAPP: Federated Test-Time Prompt Tuning for Vision-Language Models with Reliability-Aware Prompt Personalization

Hugging Face Daily Papers5d ago

Sparse Autoencoders Are a Shared Coordinate System for Symbolic Matching

What LLMs Do in the Shadows: Reading Context Reliance from the Unembedding Dark Space

QuAD: Quantitative Assessment via Distribution-Alignment for Qualitative Data Quality Enabled by a Content-Aware Multi-Agent System

From Adults to Children: Benchmarking Foundation Models for Chest X-Ray Classification under Domain Shift

UniField++: A Unified Agentic Framework with Loop Reasoning for MRI Field-Strength Enhancement

ReState: Noise-Level-Aware State Correction for Training-Free Object Removal in Few-Step Diffusion Models

Data-Efficient ConvNeXt Pretraining for High-Resolution Sentinel-1 GRD SAR Representation Learning

AI Models Can Provably Hide Arbitrary Capabilities

Is a Linear Probe Evidence of a Linear Representation?

Flat Loss, Evolving Features: Grokking in Modular Addition

FlashAttention with Statistics: Low-Overhead Attention Features for Compact Hallucination Detectors

Learning the Shared Latent Structure of Cryo-EM Inverse Problems

A Plausible Explanation Is Enough to Mislead Multimodal Judges

Spectral-Spatial Interpretation

LEMON: Equivariant Learning over Attention Maps for Zero-Shot Misbehavior Monitoring on New LLMs

GraphBEAT: Graph-based Bucketized Event Aggregation Over Time for Early Risk Stratification of Gastrointestinal Cancers

Native Geometry vs. Linear Recoverability: How Alignment Changes Platonic Scaling

Broad and Efficient PII Detection with Unified Labels

ENFR: Evidence Networks with Flow-Matched References for Amortized Simulation-Based Model Inference

Explainable Machine Learning Analysis of Temperature Effects on Mechanical Strength of Fiber-Reinforced High-Performance Concrete

ToMStories: Interpreting Theory-of-Mind in Synthetic Narratives

Semantic Latent Alignment for Flow Matching with Invertible Bi-Lipschitz Adapters

Denoising Score Matching in Diffusion Models: The Effect of Rank Constraints on Generalization

Alignment of visual brain encoding models with individual perceptual similarity judgments

OrionMoE-v1.0: A Mixture-of-Experts Foundation Model for Tabular In-Context Learning

Representation Redistribution in Sparse Autoencoders

Purify Before You Align: Improving Representation Alignment for Diffusion Models

Deep Learning for Deepfake Image Detection: Comparative Model Evaluation and Gradient-Based Visual Attribution

MAEM: Mitigating Recency Bias via a Replay-Free, Capacity-Controlled Dynamic Mixture of Experts for Class-Incremental Graph Learning

Sparse Transformer: a Controlled Testbed for Mechanistic Interpretability

CARE-SD: Explainable Reasoning Architecture for Local SLM Self-Disclosure Analysis in Textual Data