Search: “arxiv-cs-ai”

arXiv:2605.27599v1 Announce Type: cross Abstract: Agentic AI workloads - where a single user goal triggers multi-step orchestration, tool calls, retries, and failure recovery - are being targeted for

local-aiarxiv-cs-ai

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

The Ethics of LLM Sandbox and Persona Dynamics

DGX agent

arXiv:2605.28647v1 Announce Type: new Abstract: It is well known that LLM guardrails and trained persona dynamics can produce a reality gap: the distance between the world a LLM is permitted or shaped

safetyarxiv-cs-ai

28 May 2026

The Fragility of Chain-of-Thought Monitoring Across Typologically Diverse Languages

arXiv:2605.27901v1 Announce Type: cross Abstract: Chain-of-thought (CoT) monitoring has been proposed as a promising safety mechanism for detecting misaligned behavior in large language models. Howeve

The Future of Facts: Tracing the Factual Generation-Verification Gap

arXiv:2605.27564v1 Announce Type: cross Abstract: Language models are becoming the default interface to factual knowledge, yet they often verify outputs more reliably than they generate them. This gen

The Grammar of Transformers: A Systematic Review of Interpretability Research on Syntactic Knowledge in Language Models

arXiv:2601.19926v2 Announce Type: replace-cross Abstract: We present a systematic review of 337 articles evaluating the syntactic abilities of Transformer-based language models (TLMs), reporting on ov

The Illusion of Opting in AI-Mediated Consequential Decisions

arXiv:2605.28210v1 Announce Type: new Abstract: Drawing on Ullmann-Margalit's concept of opting (transformative, irrevocable, and shadowed by foreclosed alternatives), we show that current AI systems

The Importance of Being Statistically Earnest: A Critical Re-evaluation of GSM-Symbolic

arXiv:2605.28700v1 Announce Type: new Abstract: The GSM-Symbolic benchmark (Mirzadeh et al., 2025) reported consistent performance drops across 25 Large Language Models (LLMs) when tested on template-

The Obfuscation Atlas: Mapping Where Honesty Emerges in RLVR with Deception Probes

arXiv:2602.15515v2 Announce Type: replace-cross Abstract: Training against white-box deception detectors has been proposed as a way to make AI systems honest. However, such training risks models learn

The Optimal Sample Complexity of Linear Contracts

arXiv:2601.01496v2 Announce Type: replace-cross Abstract: In this paper, we settle the problem of learning optimal linear contracts from data in the offline setting, where agent types are drawn from a

The Point, the Vision and the Text: Does Point Cloud Boost Spatial Reasoning of Large Language Models? A Bias-Controlled Study

arXiv:2504.04540v2 Announce Type: replace-cross Abstract: 3D Large Language Models (LLMs) leveraging spatial information in point clouds for 3D spatial reasoning attract great attention. Despite some

Tutorials

The Principles of Diffusion Models

arXiv:2510.21890v2 Announce Type: replace-cross Abstract: This book presents the core principles that have guided the development of diffusion models, tracing their origins and showing how diverse for

tutorialsarxiv-cs-ai

The Script is All You Need: An Agentic Framework for Long-Horizon Dialogue-to-Cinematic Video Generation

arXiv:2601.17737v3 Announce Type: replace-cross Abstract: Recent advances in video generation have produced models capable of synthesizing stunning visual content from simple text prompts. However, th

Local Ai

The Shape of Overthinking: Backtracking Bursts in Long Reasoning Traces

arXiv:2605.27965v1 Announce Type: new Abstract: Reasoning models often generate long traces in which useful self-correction and unproductive revision are hard to distinguish. We study this distinction

local-aiarxiv-cs-ai

The Shape of Reasoning: Topological Analysis of Reasoning Traces in Large Language Models

arXiv:2510.20665v3 Announce Type: replace Abstract: Evaluating the quality of reasoning traces from large language models remains understudied, labor-intensive, and unreliable: current practice relies

The Well-Tempered Classifier: Some Elementary Properties of Temperature Scaling

arXiv:2602.14862v2 Announce Type: replace-cross Abstract: Temperature scaling is a simple method that allows to control the uncertainty of probabilistic models. It is mostly used in two contexts: impr

Thermodynamic properties of chemically disordered compounds via AI-driven estimation of partition function with the PULSE method

arXiv:2605.28594v1 Announce Type: cross Abstract: In this article, we present an improved version of the PULSE method (Partition function Unsupervised Learning Sampling and Evaluation) for estimating

Thinking as Compression: Your Reasoning Model is Secretly a Context Compressor

arXiv:2605.28713v1 Announce Type: new Abstract: Context compression aims to shorten long context inputs with minimal information loss for LLM inference acceleration. While existing methods have shown

Token Optimization Strategies for LLM-Based Oracle-to-PostgreSQL Migration

arXiv:2605.28557v1 Announce Type: cross Abstract: LLMs are increasingly used for software modernization, code translation, and database migration. However, LLM-based Oracle2PostgreSQL migration remain

Tool Forge: A Validation-Carrying Toolchain for Governed Agentic Execution

arXiv:2605.28000v1 Announce Type: cross Abstract: Large language model agents are increasingly expected to perform operational work: calling APIs, manipulating files, assembling workflows, and acting

Towards automated data analysis: A guided framework for LLM-based risk estimation

arXiv:2603.04631v2 Announce Type: replace Abstract: Large Language Models (LLMs) are increasingly integrated into critical decision-making pipelines, a trend that raises the demand for robust and auto

Towards Faithful Agentic XAI: A Verification Method and an Open-World Benchmark for Better Model Faithfulness

arXiv:2605.27879v1 Announce Type: new Abstract: Explainable AI (XAI) helps users interpret model behavior and identify potential faults. Agentic XAI systems use Large Language Models (LLMs) to make ex

Towards Reliable Multilingual LLMs-as-a-Judge: An Empirical Study

arXiv:2605.28710v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used for the automatic evaluation of generated text, yet most prior work focuses on English. Despite the

TRACER: Turn-level Regret Matching with Inner Reinforcement Credit for Cooperative Multi-LLM Reasoning

arXiv:2605.28699v1 Announce Type: new Abstract: Large language models increasingly rely on either reinforcement learning or multi-agent prompting to improve reasoning, yet these two paradigms remain d

Training Stratigraphy: Persistent Behavioral Artifacts in Large Language Models Observed Through Longitudinal AI-Human Interaction

arXiv:2605.28102v1 Announce Type: new Abstract: Large language models trained with Reinforcement Learning from Human Feedback (RLHF) and Constitutional AI exhibit persistent behavioral patterns that s

Transferable Reinforcement Learning via Probabilistic Latent Embeddings and Dynamic Policy Adaptation for Sim-to-Real Deployment

arXiv:2605.27659v1 Announce Type: cross Abstract: Due to limited resources and public safety concerns, deep reinforcement learning (RL) agents for many cyber-physical systems (e.g., autonomous vehicle

Tree of Thoughts as a Classical Heuristic Search Problem: Formal Foundations and Design Patterns

arXiv:2605.28566v1 Announce Type: new Abstract: Large Language Models (LLMs) have demonstrated remarkable reasoning capabilities, yet their standard generation process -- auto-regressive token predict

Trinity: Unifying Class-Agnostic Terrain and Semantic Segmentation for Unstructured Outdoor Environments by Leveraging Synthetic Data

arXiv:2605.27644v1 Announce Type: cross Abstract: Terrain understanding is fundamental for mobile robots operating in unstructured outdoor environments. Existing vision-based traversability estimation

Turning Video Models into Generalist Robot Policies

arXiv:2605.27817v1 Announce Type: cross Abstract: Video generative models have emerged as a promising robotics backbone, capable of generating videos that depict the completion of complex tasks across

Understanding Automated Program Repair Agents Through the Lens of Traceability: An Empirical Study

arXiv:2506.08311v2 Announce Type: replace-cross Abstract: Automated Program Repair (APR) agents leverage Large Language Models (LLMs) to autonomously diagnose and fix software bugs through reasoning,

Unified Synthesis of Compositional Speech and Sound from Free-Form Text Prompts

arXiv:2605.28063v1 Announce Type: cross Abstract: Audio generation has made significant progress, yet synthesizing unified audio where speech and sounds are naturally composited remains a challenge. C

UniMaia: Steering Chess Policies with Language for Human-like Play

arXiv:2605.27767v1 Announce Type: cross Abstract: Recent advances in large language models have enabled natural language to serve as a flexible interface for controlling complex systems, but often at

Unlocking Fine-Grained and Within-Utterance Speaking Style Control in Prompt-Based Text-to-Speech Models

arXiv:2605.27376v1 Announce Type: cross Abstract: While prompt-based text-to-speech (TTS) models enable natural language-driven speaking style control, they often provide limited fine-grained control

UserHarness: Harnessing User Minds for Stronger Agent Theory-of-Mind

arXiv:2605.27721v1 Announce Type: cross Abstract: Understanding what a user believes and intends is central to building effective agent assistants. This ability is often evaluated through Theory-of-Mi

Using Zero-Shot LLM-Generated Survey Data for Geographically Explicit Population Synthesis

arXiv:2605.27401v1 Announce Type: cross Abstract: There is a growing interest in utilizing synthetic populations for a diverse range of applications. At the same time, we are witnessing a tremendous g

Utility-Aware Multimodal Contrastive Learning for Product Image Generation

arXiv:2605.28733v1 Announce Type: new Abstract: Product images strongly influence consumer decision-making in online marketplaces. Empowered by multimodal contrastive learning, generative AI can outpu

VCap: Hypergeometric Rewards for Weak-to-Strong Visual Captioning

arXiv:2605.28023v1 Announce Type: cross Abstract: Visual captioning requires models to capture visual content faithfully while minimizing both omission and hallucination. As the dominant paradigm for

Verifiable Benchmarking of Long-Horizon Spatial Biology

arXiv:2605.28065v1 Announce Type: new Abstract: AI agents are increasingly useful for biological data analysis, but existing benchmarks mostly test broad biological knowledge, executable workflows, or

Verified Misguidance: Measuring Structural Citation Failures in Search-Augmented LLMs

arXiv:2605.28565v1 Announce Type: cross Abstract: Users of search-augmented LLMs rely on citations as evidence that responses are grounded in real sources, and rarely verify the cited pages themselves

VeriTrip: A Verifiable Benchmark for Travel Planning Agents over Unstructured Web Corpora

arXiv:2605.28683v1 Announce Type: new Abstract: Existing benchmarks have laid the foundation for travel planning agents by establishing API-centric paradigms. However, as the capabilities of Autonomou

VibeSearchBench: Benchmarking Long-horizon Proactive Search in the Wild

arXiv:2605.27882v1 Announce Type: cross Abstract: LLM-based agents score well on search benchmarks, yet real users consistently find results unsatisfying, revealing a persistent evaluation-experience

VidPrism: Heterogeneous Mixture of Experts for Image-to-Video Transfer

arXiv:2605.28229v1 Announce Type: cross Abstract: With the rapid development of pre-training technologies, adapting large-scale Vision-Language Models (VLMs) for video understanding ie image-to-video

Visualizing Latent Phase Structures in Locomotion Policies: A Multi-Environment Study with Temporal Feature Extension

arXiv:2605.28186v1 Announce Type: cross Abstract: Deep reinforcement learning (DRL) has been shown to achieve high performance on locomotion control tasks in MuJoCo benchmarks such as HalfCheetah, Ant

VITAL: Visual-Semantic Dual Supervision for Enhanced and Interpretable Latent Reasoning in Medical MLLMs

arXiv:2605.28422v1 Announce Type: cross Abstract: Latent reasoning enables reasoning over continuous hidden states rather than explicit tokens, avoiding the language bottleneck and inference overhead

Voluntary Collusion with Secret Tools in Competing LLM Agents

arXiv:2605.27593v1 Announce Type: new Abstract: Even when a tool is explicitly described as unfair and harmful to others, ostensibly safety-aligned LLM agents still voluntarily engage in secret collus

VULPO: Context-Aware Vulnerability Detection via On-Policy LLM Optimization

arXiv:2511.11896v3 Announce Type: replace-cross Abstract: Large language models (LLMs) have recently shown strong potential in vulnerability detection (VD). However, accurately detecting vulnerabiliti

When Context Flips, Safety Breaks: Diagnosing Brittle Safety in Aligned Language Models

arXiv:2605.27851v1 Announce Type: new Abstract: Safety benchmark scores provide incomplete evidence of deployment readiness: aligned language models often adhere to rigid rules even when a situational

When Does Memory Help Multi-Trajectory Inference for Tool-Use LLM Agents?

arXiv:2605.28224v1 Announce Type: new Abstract: Multi-trajectory inference for tool-use LLM agents - generating multiple reasoning attempts and selecting among them - benefits from transferring knowle

Local Ai

When NPUs Are Not Always Faster: A Stage-Level Analysis of Mobile LLM Inference

arXiv:2605.27435v1 Announce Type: cross Abstract: Deploying large language models (LLMs) on mobile devices increasingly relies on heterogeneous execution, yet no prior study has systematically charact

local-aiarxiv-cs-ai