AI News

Agent Pets (www.producthunt.com)

Nexus-CLI – Pure Go Bridge to DeepSeek R1/V3 with Claude Code and Ollama Support (github.com)

Google Cloud Ships llm-d 1.0 With Cache-Aware Routing for TPU and GPU Fleets (www.marktechpost.com)

Using LLM for confidence scoring is a bad idea (rumple.github.io)

Major PHP Vulnerability Leaked (jackalsyn.com)

KITE: Decoupling Kinematics and Interaction for Zero-Shot Cross-Embodiment Manipulation (arxiv.org)

A Survey of Generative Categories and Techniques in Multimodal Generative Models (arxiv.org)

Synergistic Action and Style Priors for Learning Legged Locomotion (arxiv.org)

From Outcomes to Components: A Counterfactual Evaluation Framework for LLM-Proposed Repairs (arxiv.org)

Revision-Based Learning from Interactive Experience and Feedback (arxiv.org)

IndustryForge-27B: A Domain-Enhanced Multimodal Foundation Model for Industrial CAD (arxiv.org)

IDEAgent: Agentic Quality-Diversity Search for Research Idea Generation (arxiv.org)

From Debate to Decision: Conformal Social Choice for Safe Multi-Agent Deliberation (arxiv.org)

When Models Know but Can't Linearly Say: Relational Encoding Limits Linear Vulnerability Classification in Code LLMs (arxiv.org)

ValiAssist: Scaling LLM-generated Programming Feedback Validation by Demonstration (arxiv.org)

Target-Aware Curation of Hallucination Stress Tests for Large Language Models (arxiv.org)

AutoRedClaw: Automated Red-Teaming for Claw-Style Agents (arxiv.org)

MMedAgent-Plus: Scaling Multimodal Medical Agents across Modalities, Tasks and Workflows (arxiv.org)

Profile-DP: Differential Privacy for User-Controlled Profile Release in Personalized LLM Inference (arxiv.org)

Language Guidance for Supervised Vision Training: An Empirical Study of Generalization (arxiv.org)

LoWAgent: A Low-Resource Multilingual Benchmark for Web Agents (arxiv.org)

VibeTraining-Bench: Data-Trainer-Harness Co-Design for Iterative System Improvement (arxiv.org)

From Proposal to Execution: Modeling Interaction Semantics in Shared Autonomy (arxiv.org)

VLA-Rolling: Fewer Denoising Steps, Better Control via Adaptive Action Reuse (arxiv.org)

HSPR: Recovering Software-Agent Benchmark Scores via Hierarchical Posterior Inference (arxiv.org)

HABIT-Bench: Evaluating Habit Induction from Longitudinal Weak Evidence in Agent Memory (arxiv.org)

AM-Bench: A Unified Taxonomy and Evaluation Suite for Agentic Misalignment (arxiv.org)

Memorizing Who, Knowing How: Benchmarking Personal and Procedural Memory for Persistent Assistants (arxiv.org)

Recognition Is Not Grounding: Benchmarking Multilingual Financial Evidence Extraction (arxiv.org)

Not Self-Decidable: LLMs Cannot Draw the Boundary of What an Agent Verifier Can Check (arxiv.org)

Reasoning with Continuous Latent Diffusion (arxiv.org)

Diffusion Subgoal Planning for Long-Horizon Offline Goal-Conditioned Reinforcement Learning (arxiv.org)

APOLO: Automatic Prompt Optimization for Ontology Learning (arxiv.org)

Federated learning with heavy-tailed gradient noise and communication noise: a variance-reduction based algorithm (arxiv.org)

Unified Agentic Video Editing Across Levels of Complexity and Creativity (arxiv.org)

Mitigating Model Merging Conflicts via Knowledge-Aware Task Vector Purification (arxiv.org)

Jacobian Descent for Multi-Objective Optimization (arxiv.org)

Simulating, Fast and Slow: Learning Policies for Black-Box Optimization (arxiv.org)

Learning Adaptable Dexterous Skills via Hybrid Dexterous Mimicking (arxiv.org)

Neural Implicit Energy Optimization for LiDAR Mapping (arxiv.org)

WildMAE: Structure-Aware Masked Autoencoding for Wildlife Image Reconstruction (arxiv.org)

Tabular Foundation Models Can Learn Association Rules (arxiv.org)

Agents-as-Peers: Bias-Controlled LLM Evaluation as Multi-Agent Peer Assessment (arxiv.org)

Tensor computing application programming interface: An application-oriented, lightweight API for portable high-performance tensor network computations (arxiv.org)

Theorem prover Arend (arxiv.org)

Africa-Papers: A 10-Billion-Token Multi-Domain Academic Corpus for African Foundation Language Models (arxiv.org)

De-Diffusing WAMs: Action-Sufficient Direct Prediction for Robot Control (arxiv.org)

ClawMMU: Protection-Coherent Virtual Memory for Stateful LLM Agents (arxiv.org)

BAIT: Can LLMs Find What Biases a Reader? A Benchmark for Cognitive Bias Triggers in News (arxiv.org)

TorchTPU - A Hybrid Eager/Compiled Architecture for Native PyTorch Execution on TPU (arxiv.org)