model releases

Scaling of Gaussian Kolmogorov--Arnold Networks

arXiv:2604.21174v1 Announce Type: cross Abstract: The Gaussian scale parameter (epsilon) is central to the behavior of Gaussian Kolmogorov--Arnold Networks (KANs), yet its role in deep edge-based arch

SCASeg: Strip Cross-Attention for Efficient Semantic Segmentation

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

DGX agent

arXiv:2411.17061v2 Announce Type: replace Abstract: The Vision Transformer (ViT) has achieved notable success in computer vision, with its variants widely validated across various downstream tasks, in

model-releasesarxiv-cs-cv

24 Apr 2026

SCM: Sleep-Consolidated Memory with Algorithmic Forgetting for Large Language Models

arXiv:2604.20943v1 Announce Type: new Abstract: We present SCM (Sleep-Consolidated Memory), a research preview of a memory architecture for large language models that draws on neuroscientific principl

Secure LLM Fine-Tuning via Safety-Aware Probing

arXiv:2505.16737v2 Announce Type: replace-cross Abstract: Large language models (LLMs) have achieved remarkable success across many applications, but their ability to generate harmful content raises s

Seeing Isn't Believing: Uncovering Blind Spots in Evaluator Vision-Language Models

arXiv:2604.21523v1 Announce Type: cross Abstract: Large Vision-Language Models (VLMs) are increasingly used to evaluate outputs of other models, for image-to-text (I2T) tasks such as visual question a

Separable Expert Architecture: Toward Privacy-Preserving LLM Personalization via Composable Adapters and Deletable User Proxies

arXiv:2604.21571v1 Announce Type: new Abstract: Current model training approaches incorporate user information directly into shared weights, making individual data removal computationally infeasible w

Serialisation Strategy Matters: How FHIR Data Format Affects LLM Medication Reconciliation

arXiv:2604.21076v1 Announce Type: cross Abstract: Medication reconciliation at clinical handoffs is a high-stakes, error-prone process. Large language models are increasingly proposed to assist with t

SlideAgent: Hierarchical Agentic Framework for Multi-Page Visual Document Understanding

arXiv:2510.26615v3 Announce Type: replace Abstract: Retrieval-augmented generation (RAG) extends large language models (LLMs) with external knowledge, but it must balance limited effective context, re

SocraticKG: Knowledge Graph Construction via QA-Driven Fact Extraction

arXiv:2601.10003v2 Announce Type: replace Abstract: Constructing Knowledge Graphs (KGs) from unstructured text provides a structured framework for knowledge representation and reasoning, yet current L

SparseGF: A Height-Aware Sparse Segmentation Framework with Context Compression for Robust Ground Filtering Across Urban to Natural Scenes

arXiv:2604.21356v1 Announce Type: new Abstract: High-quality digital terrain models derived from airborne laser scanning (ALS) data are essential for a wide range of geospatial analyses, and their gen

Spatial Metaphors for LLM Memory: A Critical Analysis of the MemPalace Architecture

arXiv:2604.21284v1 Announce Type: new Abstract: MemPalace is an open-source AI memory system that applies the ancient method of loci (memory palace) spatial metaphor to organize long-term memory for l

Spectral Embeddings Leak Graph Topology: Theory, Benchmark, and Adaptive Reconstruction

arXiv:2604.21094v1 Announce Type: new Abstract: Graph Neural Networks (GNNs) excel on relational data, but standard benchmarks unrealistically assume the graph is centrally available. In practice, set

Spud 🥔 and DeepSeek 🐳 V4 on the same day?! Is it Christmas? Here goes our night to bring the whale up 🔨

model-releasesfireworks-ai--x

Fireworks AI announced the release or deployment of DeepSeek V4, a large language model, alongside another project or update called 'Spud' on the same day, with the team planning to work through the n

SQLyzr: A Comprehensive Benchmark and Evaluation Platform for Text-to-SQL

arXiv:2604.21214v1 Announce Type: cross Abstract: Text-to-SQL models have significantly improved with the adoption of Large Language Models (LLMs), leading to their increasing use in real-world applic

Stealthy Backdoor Attacks against LLMs Based on Natural Style Triggers

arXiv:2604.21700v1 Announce Type: cross Abstract: The growing application of large language models (LLMs) in safety-critical domains has raised urgent concerns about their security. Many recent studie

Strategic Heterogeneous Multi-Agent Architecture for Cost-Effective Code Vulnerability Detection

arXiv:2604.21282v1 Announce Type: cross Abstract: Automated code vulnerability detection is critical for software security, yet existing approaches face a fundamental trade-off between detection accur

Structured Visual Narratives Undermine Safety Alignment in Multimodal Large Language Models

arXiv:2603.21697v2 Announce Type: replace-cross Abstract: Multimodal Large Language Models (MLLMs) extend text-only LLMs with visual reasoning, but also introduce new safety failure modes under visual

StyleID: A Perception-Aware Dataset and Metric for Stylization-Agnostic Facial Identity Recognition

arXiv:2604.21689v1 Announce Type: cross Abstract: Creative face stylization aims to render portraits in diverse visual idioms such as cartoons, sketches, and paintings while retaining recognizable ide

Subject-level Inference for Realistic Text Anonymization Evaluation

arXiv:2604.21211v1 Announce Type: new Abstract: Current text anonymization evaluation relies on span-based metrics that fail to capture what an adversary could actually infer, and assumes a single dat

SurgViVQA: Temporally-Grounded Video Question Answering for Surgical Scene Understanding

arXiv:2511.03325v3 Announce Type: replace Abstract: Video Question Answering (VideoQA) in the surgical domain aims to enhance intraoperative understanding by enabling AI models to reason over temporal

Symbolic Grounding Reveals Representational Bottlenecks in Abstract Visual Reasoning

arXiv:2604.21346v1 Announce Type: new Abstract: Vision--language models (VLMs) often fail on abstract visual reasoning benchmarks such as Bongard problems, raising the question of whether the main bot

SyMTRS: Benchmark Multi-Task Synthetic Dataset for Depth, Domain Adaptation and Super-Resolution in Aerial Imagery

arXiv:2604.21801v1 Announce Type: cross Abstract: Recent advances in deep learning for remote sensing rely heavily on large annotated datasets, yet acquiring high-quality ground truth for geometric, r

Synthetic Data in Education: Empirical Insights from Traditional Resampling and Deep Generative Models

arXiv:2604.21031v1 Announce Type: cross Abstract: Synthetic data generation offers promise for addressing data scarcity and privacy concerns in educational technology, yet practitioners lack empirical

TEMA: Anchor the Image, Follow the Text for Multi-Modification Composed Image Retrieval

arXiv:2604.21806v1 Announce Type: new Abstract: Composed Image Retrieval (CIR) is an important image retrieval paradigm that enables users to retrieve a target image using a multimodal query that cons

Temporal Taskification in Streaming Continual Learning: A Source of Evaluation Instability

arXiv:2604.21930v1 Announce Type: new Abstract: Streaming Continual Learning (CL) typically converts a continuous stream into a sequence of discrete tasks through temporal partitioning. We argue that

Tesla FSD is the first AI saving lives at scale....on real roads every single day Road accidents kill 1.19 million people every year - the #…

model-releaseselon-musk--x

Tesla FSD is the first AI saving lives at scale....on real roads every single day Road accidents kill 1.19 million people every year - the #1 cause of death for ages 5–29 94%+ of crashes are caused by

The Coding Assistant Breakdown: More Tokens Please

model-releasessemianalysis

This analysis examines the token consumption and economics of coding assistants, likely comparing different AI models' efficiency and cost-effectiveness for code generation tasks. The piece probably d

The Download: supercharged scams and studying AI healthcare

model-releasesmit-tech-review

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. We’re in a new era of AI-driven scams When ChatGPT was release

The Feedback Hamiltonian is the Score Function: A Diffusion-Model Framework for Quantum Trajectory Reversal

arXiv:2604.21210v1 Announce Type: cross Abstract: In continuously monitored quantum systems, the feedback protocol of Garcia-Pintos, Liu, and Gorshkov reshapes the arrow of time: a Hamiltonian H_{meas

The First Challenge on Remote Sensing Infrared Image Super-Resolution at NTIRE 2026: Benchmark Results and Method Overview

arXiv:2604.21312v1 Announce Type: cross Abstract: This paper presents the NTIRE 2026 Remote Sensing Infrared Image Super-Resolution (x4) Challenge, one of the associated challenges of NTIRE 2026. The

The Newton–Schulz iteration coefficients optimized by DeepSeek-V4 are surprisingly strong: they effectively normalize all singular values to…

model-releasesemad-mostaque--x

The Newton–Schulz iteration coefficients optimized by DeepSeek-V4 are surprisingly strong: they effectively normalize all singular values to 1. This matches our previous intuition: a well-balanced spe

The Path Not Taken: Duality in Reasoning about Program Execution

arXiv:2604.20917v1 Announce Type: cross Abstract: Large language models (LLMs) have shown remarkable capabilities across diverse coding tasks. However, their adoption requires a true understanding of

The Recurrent Transformer: Greater Effective Depth and Efficient Decoding

arXiv:2604.21215v1 Announce Type: new Abstract: Transformers process tokens in parallel but are temporally shallow: at position t, each layer attends to key-value pairs computed based on the previous

The Root Theorem of Context Engineering

arXiv:2604.20874v1 Announce Type: cross Abstract: Every system that maintains a large language model conversation beyond a single session faces two inescapable constraints: the context window is finit

The US CFTC sues New York, accusing the state of invading its authority to regulate prediction markets by filing lawsuits against Coinbase and Gemini (Jonathan Stempel/Reuters)

Jonathan Stempel / Reuters: The US CFTC sues New York, accusing the state of invading its authority to regulate prediction markets by filing lawsuits against Coinbase and Gemini — The U.S. Commodity F

model-releasestechmeme

The week just gets better the mad men from China do it again China is hot Deepseek V4 Pro Let’s see how it is https://huggingface.co/deepsee…

model-releasesclem-delangue--x

DeepSeek V4 Pro is a new AI model release from Chinese AI company DeepSeek, announced via Hugging Face. The post expresses enthusiasm about the model's capabilities and performance, suggesting it repr

These pelicans are kind of angry looking! Left is deepseek-v4-flash, right is deepseek-v4-pro - both generated using OpenRouter via my LLM t…

model-releasessimon-willison--x

These pelicans are kind of angry looking! Left is deepseek-v4-flash, right is deepseek-v4-pro - both generated using OpenRouter via my LLM tool 🚀 DeepSeek-V4 Preview is officially live & open-sourced!

Thinking Like a Botanist: Challenging Multimodal Language Models with Intent-Driven Chain-of-Inquiry

arXiv:2604.20983v1 Announce Type: cross Abstract: Vision evaluations are typically done through multi-step processes. In most contemporary fields, experts analyze images using structured, evidence-bas

This has been the most anticipated event in OSS and it doesn't disappoint. Congratulations to @deepseek_ai team. DSV4 Pro is now available o…

model-releasestogether-ai--x

This has been the most anticipated event in OSS and it doesn't disappoint. Congratulations to @deepseek_ai team. DSV4 Pro is now available on @togethercompute and we'll be adding a lot of capacity beh

this is probably the most important piece of software of the decade next to vllm and sglang. i'm not joking.

model-releasesjeremy-howard--x

this is probably the most important piece of software of the decade next to vllm and sglang. i'm not joking. llama.cpp at 100k stars now that 90% of the code worldwide is being written by AI agents, I

This is where we are right now. And i’m not gonna lie it feels pretty magical 🧚‍♀️ Qwen3.6 27B running inside of Pi coding agent via Llama.…

model-releasesclem-delangue--x

This is where we are right now. And i’m not gonna lie it feels pretty magical 🧚‍♀️ Qwen3.6 27B running inside of Pi coding agent via Llama.cpp on the MacBook Pro For non-trivial tasks on the @huggingf

Three reasons why DeepSeek’s new model matters

model-releasesmit-tech-review

On Friday, Chinese AI firm DeepSeek released a preview of V4, its long-awaited new flagship model. Notably, the model can process much longer prompts than its last generation, thanks to a new design t

TimePre: Bridging Accuracy, Efficiency, and Stability in Probabilistic Time-Series Forecasting

arXiv:2511.18539v2 Announce Type: replace-cross Abstract: We propose TimePre, a simple framework that unifies the efficiency of Multilayer Perceptron (MLP)-based models with the distributional flexibi

To See the Unseen: on the Generalization Ability of Transformers in Symbolic Reasoning

arXiv:2604.21632v1 Announce Type: new Abstract: We investigate the ability of decoder-only transformer models to perform abstract symbolic reasoning; specifically solving propositional logic reasoning

Tool Attention Is All You Need: Dynamic Tool Gating and Lazy Schema Loading for Eliminating the MCP/Tools Tax in Scalable Agentic Workflows

arXiv:2604.21816v1 Announce Type: new Abstract: The Model Context Protocol (MCP) has become a common interface for connecting large language model (LLM) agents to external tools, but its reliance on s

Toward Efficient Membership Inference Attacks against Federated Large Language Models: A Projection Residual Approach

arXiv:2604.21197v1 Announce Type: new Abstract: Federated Large Language Models (FedLLMs) enable multiple parties to collaboratively fine-tune LLMs without sharing raw data, addressing challenges of l

Towards Multimodal Active Learning: Efficient Learning with Limited Paired Data

arXiv:2510.03247v2 Announce Type: replace-cross Abstract: Active learning (AL) is a principled strategy to reduce annotation cost in data-hungry deep learning. However, existing AL algorithms focus al

Towards Universal Tabular Embeddings: A Benchmark Across Data Tasks

arXiv:2604.21696v1 Announce Type: new Abstract: Tabular foundation models aim to learn universal representations of tabular data that transfer across tasks and domains, enabling applications such as t