Search: “models”

ESLD (External Surrogate Latent Defense): A Latent-Space Architecture for Faster, Stronger Prompt-Injection Defense

arXiv:2605.18918v1 Announce Type: cross Abstract: Modern AI assistants are agentic. To answer a single user request, the underlying language model pulls in information from many sources, such as web s

safetyarxiv-cs-ai

Model Releases

FedMental: Evaluating Federated Learning for Mental Health Detection from Social Media Data

DGX agent

arXiv:2605.18936v1 Announce Type: cross Abstract: Social media text data are often used to train Machine Learning (ML) models to identify users exhibiting high-risk mental health behaviors. However, s

model-releasesarxiv-cs-cl

20 May 2026

From Prompts to Pavement Through Time: Temporal Grounding in Agentic Scene-to-Plan Reasoning

arXiv:2605.19824v1 Announce Type: new Abstract: Recent attempts to support high-level scene interpretation and planning in Autonomous Vehicles (AVs) using ensembles of Large Language Models (LLMs) and

GRAB: A Risk Taxonomy--Grounded Benchmark for Unsupervised Topic Discovery in Financial Disclosures

model-releasesarxiv-cs-cl

arXiv:2509.21698v2 Announce Type: replace Abstract: Risk categorization in 10-K risk disclosures matters for oversight and investment, yet no public benchmark evaluates unsupervised topic models for t

Hallucination as Exploit: Evidence-Carrying Multimodal Agents

arXiv:2605.19192v1 Announce Type: new Abstract: Multimodal agents use screenshots, documents, and webpages to choose tool calls. When a false visual claim triggers a click, email, extraction, or trans

How Does Overparameterization Affect Machine Unlearning of Deep Neural Networks?

arXiv:2503.08633v2 Announce Type: replace Abstract: Machine unlearning is the task of updating a trained model to forget specific training data without retraining from scratch. In this paper, we inves

safetyarxiv-cs-lg

Local Ai

INSIGHTS: Demonstration-Based Summaries of Time Series Predictors

arXiv:2605.18849v1 Announce Type: cross Abstract: Explainability methods have progressed rapidly, but global explanations for time-series models remain underdeveloped, with most approaches focusing on

local-aiarxiv-cs-ai

Learn-by-Wire Training Control Governance: Bounded Autonomous Training Under Stress for Stability and Efficiency

arXiv:2605.19008v1 Announce Type: new Abstract: Modern language-model training is increasingly exposed to instability, degraded runs, and wasted compute, especially under aggressive learning-rate, sca

Measuring Safety Alignment Effects in Autonomous Security Agents

arXiv:2605.19722v1 Announce Type: cross Abstract: Do stock safety-aligned language models and their uncensored or abliterated derivatives behave differently when run as autonomous security agents? Sin

MSAVBench: Towards Comprehensive and Reliable Evaluation of Multi-Shot Audio-Video Generation

arXiv:2605.20183v1 Announce Type: new Abstract: Video generation is rapidly evolving from single-shot synthesis to complex multi-shot audio-video (MSAV) narratives to meet real-world demands. However,

Not All Tokens Are Worth Caching: Learning Semantic-Aware Eviction for LLM Prefix Caches

arXiv:2605.18825v1 Announce Type: new Abstract: Prefix caching is a key optimization in Large Language Model (LLM) serving, reusing attention Key-Value (KV) states across requests with shared prompt p

Local Ai

Perceptual misalignment of texture representations in convolutional neural networks

arXiv:2604.01341v2 Announce Type: replace Abstract: Mathematical modeling of visual textures traces back to Julesz's intuition that texture perception in humans is based on local correlations between

local-aiarxiv-cs-cv

✨ Personal AI is the next computing platform. AI is shifting from something you access to something you build with, locally, at the edge, an…

model-releasesclem-delangue--x

✨ Personal AI is the next computing platform. AI is shifting from something you access to something you build with, locally, at the edge, and across systems. We’re unlocking new possibilities for deve

Precision Tracked Transformer via Kalman Filtering, Kriging and Process Noise

arXiv:2605.18832v1 Announce Type: cross Abstract: The Transformer is the foundational building block of modern AI, yet offers no principled handling of uncertainty, which is prevalent in real applicat

PRISM: A Benchmark for Programmatic Spatial-Temporal Reasoning

arXiv:2605.19382v1 Announce Type: new Abstract: Programmatic video generation through code offers geometric precision and temporal coherence beyond pixel-level diffusion models, yet rigorously evaluat

Reporting from Google I/O 2026 with the four biggest themes from one of the biggest AI labs in the world. 🎤 Voice AI as an interface Google…

model-releasesallie-k--miller--x

Reporting from Google I/O 2026 with the four biggest themes from one of the biggest AI labs in the world. 🎤 Voice AI as an interface Google and Samsung announced new Gemini-powered glasses with Gentle

Robustness and Regularization in Hierarchical Re-Basin

arXiv:2510.09174v3 Announce Type: replace Abstract: This paper takes a closer look at Git Re-Basin, an interesting new approach to merge trained models. We propose a hierarchical model merging scheme

researcharxiv-cs-lg

STAR-PolyaMath: Multi-Agent Reasoning under Persistent Meta-Strategic Supervision

arXiv:2605.19338v1 Announce Type: cross Abstract: Frontier AI models and multi-agent systems have led to significant improvements in mathematical reasoning. However, for problems requiring extended, l

Streamlined Constraint Reasoning via CNN Pattern Recognition on Enumerated Solutions

arXiv:2605.19895v1 Announce Type: new Abstract: Constraint programming practitioners accelerate hard problems through a layered set of techniques applied in order of risk. Standard hardening (symmetry

Structured Layout Priors for Robust Out-of-Distribution Visual Document Understanding

arXiv:2605.19866v1 Announce Type: new Abstract: Vision-Language Models (VLMs) parse documents end-to-end but frequently break down on layouts unlike those seen in training. We attribute this to a two-

Hardware

TIDE: Efficient and Lossless MoE Diffusion LLM Inference with I/O-aware Expert Offload

arXiv:2605.20179v1 Announce Type: new Abstract: Diffusion Large Language Models (dLLMs) have emerged as a competitive alternative to autoregressive (AR) models, offering better hardware utilization an

hardwarearxiv-cs-cl

Toto 2.0: Time Series Forecasting Enters the Scaling Era

arXiv:2605.20119v1 Announce Type: cross Abstract: We show that time series foundation models scale: a single training recipe produces reliable forecast-quality improvements from 4M to 2.5B parameters.

Towards Camera-Robust 3D Localization: Equation-Anchored Tool-Use for MLLMs

arXiv:2605.19528v1 Announce Type: new Abstract: 3D localization in Multimodal Large Language Models (MLLMs), including 3D object detection and 3D visual grounding, is fundamentally limited by camera i

Towards Distillation Guarantees under Algorithmic Alignment for Combinatorial Optimization

arXiv:2605.20074v1 Announce Type: new Abstract: Distillation transfers knowledge from a large model trained on broad data to a smaller, more efficient model suitable for deployment. In structured pred

safetyarxiv-cs-lg

Trust or Abstain? A Self-Aware RAG Approach

model-releasesarxiv-cs-cl

arXiv:2605.18792v1 Announce Type: cross Abstract: Retrieval-augmented generation (RAG) improves large language models (LLMs) by incorporating external evidence, but it also introduces knowledge confli

ViroGym: Realistic Large-Scale Benchmarks for Evaluating Viral Proteins

arXiv:2603.06740v2 Announce Type: replace-cross Abstract: Protein language models (pLMs) have shown strong potential for zero-shot prediction of missense variant effects, yet systematic benchmarking o

WARC-Bench: Web Archive Based Benchmark for GUI Subtask Executions

arXiv:2510.09872v2 Announce Type: replace-cross Abstract: Training web agents to navigate complex, real-world websites requires them to master extit{subtasks} - short-horizon interactions on multiple

What Makes a Representation Good for Single-Cell Perturbation Prediction?

arXiv:2605.19343v1 Announce Type: new Abstract: Single-cell perturbation modeling is fundamental for understanding and predicting cellular responses to genetic perturbations. However, existing approac

Worldlier: native support for 48 world languages and improved efficiency in non-European languages.

Worldlier is a Cohere language model that natively supports 48 world languages with improved efficiency, particularly for non-European languages. This represents an expansion of language coverage beyo

model-releasescohere--x

Local Ai

Your Neighbors Know: Leveraging Local Neighborhoods for Backdoor Detection in Decentralized Learning

arXiv:2605.19969v1 Announce Type: new Abstract: Decentralized learning (DL) is an emerging machine learning paradigm where nodes collaboratively train models without a central server. However, the col

local-aiarxiv-cs-lg

A Machine with Short-Term, Episodic, and Semantic Memory Systems

arXiv:2212.02098v5 Announce Type: replace Abstract: Inspired by the cognitive science theory of the explicit human memory systems, we have modeled an agent with short-term, episodic, and semantic memo

A More Word-like Image Tokenization for MLLMs

arXiv:2605.17954v1 Announce Type: cross Abstract: Modern multimodal large language models (MLLMs) typically keep the language model fixed and train a visual projector that maps the pixels into a seque

researcharxiv-cs-ai

Auditing Multimodal LLM Raters: Central Tendency Bias in Clinical Ordinal Scoring

arXiv:2605.16386v1 Announce Type: new Abstract: Multimodal large language models (LLMs) are increasingly explored as automated evaluators in clinical settings, yet their scoring behavior on ordinal cl

BESplit: Bias-Compensated Split Federated Learning with Evidential Aggregation

arXiv:2605.17508v1 Announce Type: cross Abstract: Split Federated Learning (SFL) enables privacy-preserving collaborative training by partitioning models between clients and a server. However, under n

Beyond Point-Wise Matching: Structural Representation Alignment for Accelerating Diffusion Transformers

arXiv:2605.16949v1 Announce Type: new Abstract: Recent advances in Diffusion Transformers (DiTs) demonstrate that aligning noisy latent states with well-trained semantic features-as pioneered by Repre

Boundedly Rational Meta-Learning in Sequential Consumer Choice

arXiv:2605.16532v1 Announce Type: new Abstract: Many consumer decisions are repeated choices under uncertainty. Standard models capture these decisions using Bayesian learning and dynamic programming:

CADS: Conformal Adaptive Decision System for Cost-Efficient Image Classification

arXiv:2605.16401v1 Announce Type: new Abstract: While high-capacity AI models have advanced state-of-the-art performance, their practical deployment is often hindered by high inference costs, environm

researcharxiv-cs-cv

CAM-Bench: A Benchmark for Computational and Applied Mathematics in Lean

arXiv:2605.17255v1 Announce Type: new Abstract: Formal theorem-proving benchmarks enable mechanically verifiable evaluation of mathematical reasoning in large language models. However, existing benchm

Can LLMs Refuse Questions They Do Not Know? Measuring Knowledge-Aware Refusal in Factual Tasks

arXiv:2510.01782v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) should refuse to answer questions beyond their knowledge. This capability, which we term knowledge-aware refusal,

researcharxiv-cs-ai

CasualSynth: Generating Structurally Sound Synthetic Data

arXiv:2605.17528v1 Announce Type: cross Abstract: Large Language Models (LLMs) generate realistic synthetic data but offer no guarantee that their outputs respect the causal mechanisms governing the t

ClaHF: A Human Feedback-inspired Reinforcement Learning Framework for Improving Classification Tasks

arXiv:2605.17458v1 Announce Type: new Abstract: Text classification models are typically trained via supervised fine-tuning (SFT). However, SFT essentially performs behavior cloning from instance-wise

safetyarxiv-cs-lg

Closing the Gap at CRAC 2026: Two-Stage Adaptation for LLM-Based Multilingual Coreference Resolution

model-releasesarxiv-cs-cl

arXiv:2605.16984v1 Announce Type: new Abstract: We present our submission to the LLM track of the 2026 Computational Models of Reference, Anaphora and Coreference (CRAC 2026) shared task. With an aver

ContractBench: Can LLM Agents Preserve Observation Contracts?

arXiv:2605.17281v1 Announce Type: cross Abstract: Tool-augmented LLM agents call APIs whose intermediate outputs, such as presigned URLs, session tokens, and OAuth state parameters, are observation co

CPMobius: Iterative Coach-Player Reasoning for Data-Free Reinforcement Learning

arXiv:2602.02979v2 Announce Type: cross Abstract: Large Language Models (LLMs) have demonstrated strong potential in complex reasoning, yet their progress remains fundamentally constrained by reliance

DBES: A Systematic Benchmark and Metric Suite for Evaluating Expert Specialization in Large-Scale MoEs

arXiv:2605.18498v1 Announce Type: cross Abstract: Expert specialization in Mixture-of-Experts (MoE) models remains poorly understood, with traditional evaluations conflating architectural load-balanci

Disappointing pricing trend with Gemini 3.5 Flash. 22.5x pricier than 2.0 Flash which came out 15 months ago (9.00 vs 0.40). Are Flash mod…

model-releasesjeremy-howard--x

Disappointing pricing trend with Gemini 3.5 Flash. 22.5x pricier than 2.0 Flash which came out 15 months ago (9.00 vs 0.40). Are Flash models supposed to get this much more expensive, or is Pro just b

Distributed Perceptron under Bounded Staleness, Partial Participation, and Noisy Communication

arXiv:2601.10705v3 Announce Type: replace Abstract: We study a semi-asynchronous client-server perceptron trained via iterative parameter mixing (IPM-style averaging): clients run local perceptron upd

DriveSafer: End-to-End Autonomous Driving with Safety Guidance

arXiv:2605.16737v1 Announce Type: cross Abstract: End-to-End (E2E) autonomous driving models have shown growing capability in recent years, with performance improving on increasingly challenging bench