All entries

arXiv:2608.26449v1 Announce Type: new Abstract: Byte-level BPE tokenizers that use the HuggingFace ByteLevel pre-tokenizer inherit GPT-2's word regex, where a word is defined as p{L}+, one or more Uni

tutorialsarxiv-cs-cl

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

WALL-SS: Scaling Long-horizon World Models via Next-Scale Autoregression

DGX agent

arXiv:2608.26239v1 Announce Type: new Abstract: Generative world models provide robots with predictive models of how the world evolves under interaction, with growing potential for simulation, plannin

safetyarxiv-cs-ro

28 Aug 2026

Agents

What Makes Good Agentic Data? An ACE Lens on Data Generation for LLM Agents

arXiv:2608.27260v1 Announce Type: new Abstract: LLM agents increasingly rely on generated interaction data to learn how to interact with external environments. Agentic data generation must maintain co

agentsarxiv-cs-ai

What the 'Spotless' Mind Remembers: How Knowledge Entanglement Shapes What Leaks After Unlearning in LLMs

arXiv:2510.25732v2 Announce Type: replace-cross Abstract: Unlearning in large language models (LLMs) is usually evaluated as whether an 'unlearned' fact can be recovered. We instead ask whether a fact

researcharxiv-cs-ai

When Interference Graphs Evolve: Doubly Robust Estimation of Dynamic Peer Effects

arXiv:2608.27187v1 Announce Type: new Abstract: Peer effects are difficult to estimate when interaction graphs evolve because pre-assignment network history, dynamic peer exposure, and post-assignment

When Is Noise Response Universal? Tokenization as the Hidden Variable in Language Models

arXiv:2608.26319v1 Announce Type: new Abstract: The performance of textual neural models often degrades when their inputs are corrupted by noise such as typos, OCR errors, or dropped words. We study t

When Is the Sharp Covariance Envelope Tight? Feature-Only Geometry for Volume-Sampled Least Squares

arXiv:2608.26877v1 Announce Type: new Abstract: Prior analyses by Derezinski and Warmuth established all-size sampling identities, selected-OLS unbiasedness, and inverse moments for ordinary volume sa

applicationsarxiv-cs-lg

When Memory Takes Gradients: Collaborative Vector Memory for Agentic Recommender Systems

arXiv:2608.26895v1 Announce Type: cross Abstract: Agentic recommender systems ground each decision of a large language model (LLM) in a persistent memory of the user, and in existing agents that memor

safetyarxiv-cs-ai

When Privacy Hurts Mergeability: Geometry-Aware Model Merging under Differential Privacy

arXiv:2608.26655v1 Announce Type: new Abstract: Model merging promises to construct a single multi-task model from independently fine-tuned task models without accessing the original task data. This m

When Text Misleads: Inconsistent-Aware Reasoning for Audio-Grounded Dialogue

arXiv:2608.27176v1 Announce Type: cross Abstract: Understanding spoken dialogue requires joint reasoning over lexical content and paralinguistic acoustic signals such as emotion and conversational int

When the Canonical Completion Is Wrong: Formalizing and Measuring the Jump in Large Language Models

arXiv:2608.26187v1 Announce Type: cross Abstract: Whether large language models (LLMs) can perform the abductive leap from evidence to a new system of axioms, commonly referred to as a jump, has recen

researcharxiv-cs-ai

Agents

When Tool Outputs Become Commands: Separating Action Induction from Runtime Authorization in Tool-Augmented LLM Agents

arXiv:2608.27146v1 Announce Type: new Abstract: Tool-augmented LLM agents must rely on untrusted runtime Observations to complete open-ended tasks; however, when tool outputs no longer merely provide

agentsarxiv-cs-ai

Which India Survives Translation? Narrative Homogenisation Across Indian Oral Traditions in LLMs

arXiv:2608.26123v1 Announce Type: new Abstract: Large language models (LLMs) are trained predominantly on English-language internet text that over-represents certain cultural narratives, raising conce

Which Metrics Save the Most Human Annotation? Prediction-Powered Evaluation and Meta-Evaluation

arXiv:2608.26638v1 Announce Type: new Abstract: Across various non-verifiable tasks, human evaluation is reliable but expensive, while automatic metrics are more scalable but often biased. Building on

Who Remains, What Changes: Identity Anchored Composed Gait Retrieval

arXiv:2608.26632v1 Announce Type: new Abstract: Gait recognition has achieved remarkable progress, yet existing methods remain confined to rigid visual matching and often overlook the potential of nat

researcharxiv-cs-cv

Why Current XAI Is Not Enough for Arabic NLP: A Critical Survey of the Explainability Gap

arXiv:2608.26144v1 Announce Type: new Abstract: Explainable AI (XAI) is now a major theme in NLP; however, Arabic NLP remains under-explained in three connected senses. First, there is a method gap: A

Why did My Robot Just Change Personality? Prompting Guidelines for a Grounded Robot Persona in LLM-Based HRI

arXiv:2608.26182v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used for verbal interaction in social robots, yet prompt design in human-robot interaction (HRI) remains u

safetyarxiv-cs-ai

Why not to use the Gaussian kernel

arXiv:2608.26974v1 Announce Type: cross Abstract: Kernels measure similarity or correlation in tasks such as regression and classification. The Gaussian kernel, other names of which include squared ex

Why RAGs Hallucinate: Penalty-Aware Evaluation of Retrieval-Augmented Generation Systems with Knowledge-Gap Canaries

arXiv:2608.26385v1 Announce Type: cross Abstract: Volume-based accuracy rewards retrieval-augmented generation (RAG) systems for guessing: a system that answers everything outscores one that declines

WikiSkill: Compiling Agent Experience into Persistent Knowledge for Skill Evolution

arXiv:2608.27454v1 Announce Type: new Abstract: Agent skills package specialized knowledge and workflows into reusable resources that extend AI agent capabilities. Recent work automatically discovers

Your Voice Cloning System is Secretly a Voice Anonymizer

arXiv:2608.27360v1 Announce Type: new Abstract: Speaker anonymization suppresses speaker-identifying attributes from speech while preserving linguistic content and quality. We propose repurposing XTTS

Zero-Shot Self-Orchestration with Ledger-Based Control for Improved LLM Coding Performance

arXiv:2608.26480v1 Announce Type: cross Abstract: Multi-agent large language model systems are widely reported to beat single-model baselines, but the evidence is mixed, and comparisons are usually co

Zero-Shot Video Restoration and Enhancement with Text-to-Image Latent Diffusion Models and Multi-Modal References

arXiv:2608.26476v1 Announce Type: new Abstract: Zero-shot image restoration methods with text-to-image latent diffusion models have achieved great success in universal image restoration tasks without

researcharxiv-cs-cv

3DGAA: Realistic and Robust 3D Gaussian-based Adversarial Attack for Autonomous Driving

arXiv:2507.09993v4 Announce Type: replace Abstract: Camera-based perception in connected and autonomous vehicles remains exposed to physical adversarial attacks. Prior attacks often either optimize im

safetyarxiv-cs-cv

3DGS-HPC: Distractor-free 3D Gaussian Splatting with Hybrid Patch-wise Classification

arXiv:2603.07587v2 Announce Type: replace Abstract: 3D Gaussian Splatting (3DGS) has demonstrated remarkable performance in novel view synthesis and 3D scene reconstruction, but its quality often degr

applicationsarxiv-cs-cv

4DGS-WAM: Bridging Past and Future with an Object-Centric World Action Model based on 4D Gaussian Splatting

arXiv:2608.25956v1 Announce Type: new Abstract: Current world action models (WAMs) typically operate on 2D visual data. These models can achieve exceptional visual quality, but they lack explicit spat

safetyarxiv-cs-cv

4DStreamCtrl: Interactive Video Generation with Online 4D Control

model-releasesarxiv-cs-cv

arXiv:2608.25479v1 Announce Type: new Abstract: Generative video models now synthesize footage nearly indistinguishable from reality. Their promise as interactive tools hinges on fine-grained control

A Comedy of Estimators: On KL Regularization in RL Training of LLMs

arXiv:2512.21852v4 Announce Type: replace Abstract: The reasoning performance of large language models (LLMs) can be substantially improved by training them with reinforcement learning (RL). The RL ob

A Constitutive Markov Physics-Informed Neural Operator (MPNO) for Autoregressive Stability in Transient Dynamics

arXiv:2608.25744v1 Announce Type: new Abstract: Neural operators applied to transient-dynamics PDEs with strong discontinuities exhibit autoregressive instability: in concrete-penetration stress-field

A Dual-Transformer for Multi-Camera View Recommendation

arXiv:2608.25601v1 Announce Type: new Abstract: Multi-camera systems are foundational to modern media production, and multi-camera editing is a critical task. This involves the proper selection of the

applicationsarxiv-cs-cv

A General-Purpose Framework for Chemical Reaction Representation with Atomic Correspondence and Flexible Condition Adaptation

arXiv:2411.17629v3 Announce Type: replace Abstract: Motivation: Organic synthesis is fundamental to the chemical industry, particularly in domains such as pharmaceutical development. While artificial

A General-Purpose Molecular Foundation Model Transfers Across Diverse Olfactory Tasks

arXiv:2608.25893v1 Announce Type: new Abstract: Foundation models have transformed molecular property prediction, yet it remains unclear whether a molecular foundation model, fine-tuned on a single ca

A Hierarchical Synergistic Deep Learning Framework Integrating Composition, Structure, and Ionic Transport for Solid-State Electrolyte Discovery

arXiv:2608.25592v1 Announce Type: cross Abstract: Inorganic solid-state electrolytes must combine high room-temperature ionic conductivity, a wide electrochemical window, excellent electronic insulati

A Layered Security Framework Against Prompt Injection in RAG-Based Chatbots

arXiv:2606.19660v2 Announce Type: replace-cross Abstract: Prompt injection is ranked as the most critical vulnerability in large language model (LLM) deployments by the OWASP Top 10 for LLM Applicatio

Local Ai

A Lightweight Multimodal Vision-Language Framework for Early-Stage Anatomical Green Fruit Classification in Commercial Orchards

arXiv:2608.24935v1 Announce Type: new Abstract: Accurate identification of early-stage apple fruitlet anatomical structures, including the calyx, fruitlet body, and peduncle, is essential for robotic

local-aiarxiv-cs-cv

A meta-algorithm for ab initio reconstruction of complex mixtures in cryo-EM

arXiv:2608.25388v1 Announce Type: cross Abstract: We describe a systematic approach for spawning and aggregating multi-class cryo-EM reconstruction jobs. This approach formalizes standard ad hoc strat

A Multi-View Coupled Tensor Decomposition for Lightweight Online Adaptive Traffic Prediction

arXiv:2608.25498v1 Announce Type: cross Abstract: Accurate online traffic prediction is essential for intelligent transportation systems, where forecasting must be performed continuously under imperfe

applicationsarxiv-cs-lg

A Primer on Computational Semantics for Artificial Intelligence Systems

arXiv:2608.25022v1 Announce Type: new Abstract: As people adopt transformer-based language models (e.g., ChatGPT and Gemini) for an increasing number of use-cases, it is important to know how such mod

A Self-Evolving Multi-Agent Framework Defense against LLM Jailbreak Attacks

arXiv:2608.26008v1 Announce Type: cross Abstract: Large language models (LLMs) remain vulnerable to jailbreak attacks that exploit techniques such as role-playing, obfuscation, code transformation, an

A Statistical Audit of Physical AI Benchmark Redundancy

model-releasesarxiv-cs-ro

arXiv:2608.25940v1 Announce Type: new Abstract: Physical AI models are evaluated on suites of benchmarks that differ across model reports, leaving the model-by-benchmark matrix sparse and the relation

A Storage-Retrieval Gap in Parametric Knowledge Graph Memory

arXiv:2608.25489v1 Announce Type: cross Abstract: Graph retrieval-augmented generation places retrieved subgraphs into the model's context window at query time, paying a recurring token cost and expos

A Taxonomy of Construction Task Activities for Robot Workers

arXiv:2608.25395v1 Announce Type: new Abstract: Recent vision-language-action models offer a path toward robots with broader repertoires than conventional task-specific systems. Construction deploymen

researcharxiv-cs-ro

A Tendon-Driven Five-Fingered Hand with Distributed Tactile Perception for Dexterous Manipulation

arXiv:2608.25547v1 Announce Type: new Abstract: To apply the techniques of embodied artificial intelligence to human-oid robots for complex manipulations, dexterous robotic hands are indispensable, wh

researcharxiv-cs-ro

A Token-Level Analysis of Sampled-Token Reverse-KL On-Policy Distillation

arXiv:2608.25643v1 Announce Type: cross Abstract: On-policy distillation (OPD) supervises a student on its own trajectories with token-level signals from a frozen teacher, yet how a sampled loss alloc

safetyarxiv-cs-cl

A Visual Dependence-Aware Framework for Multimodal Unsupervised Continual Post-Training

arXiv:2608.26095v1 Announce Type: new Abstract: In this paper, we explore a novel task of Multimodal Unsupervised Continual Post-Training (MU-CPT), enabling deployed MLLMs to continually evolve from s

safetyarxiv-cs-cv

Activation-Space Order-Swap Geometry: A Site-Asymmetry Audit

arXiv:2608.25315v1 Announce Type: new Abstract: Order-dependent activation statistics are often interpreted as evidence of interaction, but that interpretation can be confounded by where interventions

Activation Steering Transfer to Agents: One Gain Ratio Does Not Identify Potency and Efficacy

arXiv:2607.09156v2 Announce Type: replace Abstract: Additive activation steering is calibrated in single-turn chat and then deployed inside agent scaffolds. The quantity usually reported for that move

safetyarxiv-cs-lg

Adaptive Bayes exactly tracks information over intrinsic time

arXiv:2607.08789v2 Announce Type: replace Abstract: Bayesian and multiplicative-weights updates reweight experts, models, or actions from sequential feedback. We show that the regret of any such updat