Search: “safety”

One Frozen Simulator Is Not Enough: Simulator Collapse in Multi-Agent RL

arXiv:2608.12253v1 Announce Type: cross Abstract: Multi-agent reinforcement learning for human-AI interaction typically relies on a single large language model to simulate user behavior. We show that

PAIR: Pairwise-Aware Inclusion Reweighting for Adaptive Rollout Allocation in RLVR

DGX agent

arXiv:2608.11368v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) spends most of its compute generating groups of long reasoning trajectories. Recent allocators red

safetyarxiv-cs-lg

13 Aug 2026

Policy-as-logic for robust reasoning over rules

arXiv:2608.11905v1 Announce Type: new Abstract: In many practical applications of generative AI systems, from tax rules to airline baggage allowance, responses to natural language queries must respect

Policy-Induced Hand Priors in Humanoid Dual-Arm Manipulation: Diagnosing and Mitigating Initial-Pose Dependence

arXiv:2608.11769v1 Announce Type: new Abstract: Vision-language-action (VLA) policies are expected to operate robustly across variations in the robot's initial configuration, yet aggregate task succes

safetyarxiv-cs-ro

Post-Training with Policy Gradients: Optimality and the Base Model Barrier

arXiv:2603.06957v2 Announce Type: replace-cross Abstract: We study post-training linear autoregressive models with outcome and process rewards. Given a context oldsymbol{x}, the model must predict the

Prompt-Driven Exploration

arXiv:2607.08837v2 Announce Type: replace-cross Abstract: Exploration is essential to RL since a policy cannot improve by repeatedly sampling the behaviors it already prefers. Standard methods inject

Proportional Committee Elections with Positive and Negative Votes

arXiv:2503.01985v2 Announce Type: replace-cross Abstract: In the classic committee election setting each voter approves a subset of candidates and the goal is to select k winners based on these prefer

RA-ClipScore: Making Generative Model Evaluation More Interpretable

arXiv:2608.12088v1 Announce Type: new Abstract: Generative models can produce images nearly indistinguishable from real data, yet rigorous and interpretable evaluation remains challenging. Conventiona

Redistribution-based Cost Inference Improves Sparse Safe Offline RL

arXiv:2608.12306v1 Announce Type: cross Abstract: Safe offline RL typically assumes access to dense per-step cost annotations, but in practice supervisors provide only trajectory-level stop-feedback:

REOPD: Reliability-Adaptive Reward Extrapolation for On-Policy Distillation

arXiv:2608.11698v1 Announce Type: cross Abstract: On-policy distillation (OPD) trains a student on its own trajectories under dense token-level supervision from a teacher. Reward-extrapolation methods

Reoptimization Algorithms for Contextual Bandits with Knapsack Constraints

arXiv:2608.11383v1 Announce Type: new Abstract: We study new algorithms for Contextual Bandits with Knapsack. In these problems, there are finitely many types of customers, products, and resources. Ea

Robust Multi-Tier Infant-Centered Audio Understanding with Whisper via Structured Speaker Conditioning

arXiv:2608.11587v1 Announce Type: cross Abstract: Recent advances in model design and self-supervised audio representations have improved speech and audio understanding, yet infant-centered naturalist

ScaleVid: Geometry-Aware Video Object Scaling with Mesh-Free Inference

arXiv:2608.12232v1 Announce Type: new Abstract: Geometry-aware video object scaling aims to anisotropically resize the object along object-centric axes while preserving geometric plausibility, tempora

Small Data Explainer -- The impact of small data methods in everyday life

arXiv:2507.11773v2 Announce Type: replace-cross Abstract: The emergence of breakthrough artificial intelligence (AI) techniques has led to a renewed focus on how small data settings, i.e., settings wi

Stolen LLM Reasoning: How come OpenAI, Anthrophic, Google have the same vulnerabilities?

If you haven't checked the paper: https://arxiv.org/abs/2608.09867 TLDR: the authors show that you can swap out the 'encrypted' reasoning of the biggest model, like Opus, Sol, and put them into weaker

safetyr-localllama

Through Van Gogh's Eyes: Global Style Transfer with Diffusion Mod

arXiv:2608.11546v1 Announce Type: new Abstract: Artistic image synthesis aims to recreate the expressive visual identity of a target artist, yet existing methods often fail to capture an artist's glob

ToolHazard: Scaling Adversarial Environments for Security Evaluation and Alignment of LLM-based Agents

arXiv:2608.11878v1 Announce Type: cross Abstract: Large language model (LLM) agents integrated with external tools are vulnerable to indirect prompt injections embedded in environmental states. Howeve

Toward Meaningful Transparency for AI Chatbots: Disclosing Persuasive Intent Reduces Persuasion

arXiv:2608.11794v1 Announce Type: cross Abstract: The growing role of AI-generated content and AI-enabled systems in public communication has led regulators to demand clear disclosure of content prove

Towards a Formal Definition of Agent Memory: Basis, Span, Optimality, and the Sequential Memory Problem

arXiv:2608.11654v1 Announce Type: new Abstract: Despite the wide deployment of memory in large-model agents, there is no unified formal account of what a memory is or when it is optimal. This paper ta

Towards Human Motion World Models via Executable Behaviour Representations

arXiv:2604.18064v2 Announce Type: replace Abstract: Human motion world models should capture motion's intentionality by being executable: adaptable to different actions and capable of assessing motion

Towards Understanding On-Policy Distillation through the Lens of Test-Time Scaling

arXiv:2608.11829v1 Announce Type: cross Abstract: On-policy distillation (OPD) has emerged as a promising post-training technique for enhancing LLM reasoning. It is commonly believed to enable the stu

UniSwap: Streaming Audio-Visual Identity Swapping for Talking Videos

arXiv:2608.11752v1 Announce Type: new Abstract: Talking-video character replacement requires coordinated transfer of appearance and voice while preserving the source motion, scene, linguistic content,

Variable Selection in the Context of AI Fairness

arXiv:2608.11251v1 Announce Type: cross Abstract: Fairness in AI systems has become more important with recent regulatory demands, such as the EU AI Act. Traditional approaches often do not take into

Who Would You Vote For? Auditing Political Alignment in LLMs: An Italian Case-Study

arXiv:2608.11649v1 Announce Type: new Abstract: As users increasingly turn to Large Language Models (LLMs) for information and advice on political matters, particularly during election periods, the po

Why AI Detection Fails for Academic Integrity

arXiv:2608.11256v1 Announce Type: new Abstract: Institutions use commercial AI detectors for academic integrity, yet detectors cannot distinguish AI editing from full LLM drafts and may treat both as

A Joint-Distribution Route to Fair Representations with Continuous Sensitive Attributes

arXiv:2608.10470v1 Announce Type: new Abstract: Fair representation learning with a continuous sensitive attribute S requires a representation Z that is statistically independent of S. Existing criter

Actions Speak Louder than Words: Measuring Cross-Lingual Policy Retention in Tool-Using Agents

arXiv:2608.11110v1 Announce Type: new Abstract: When a tool-using agent is given the same task in a different language, does it still take the same steps? Multilingual evaluation rarely asks: it compa

// Actions Speak Louder Than Words // Multilingual agent evaluation compares final answers and throws the trajectory away. The trajectory fi…

// Actions Speak Louder Than Words // Multilingual agent evaluation compares final answers and throws the trajectory away. The trajectory fixes cost, latency, failure mode, and auditability. New resea

safetydair-ai--x

AdvFD: Boosting Visual Generation via Adversarial Fr'echet Distance Loss

arXiv:2608.11205v1 Announce Type: new Abstract: Frechet distance has recently emerged as an effective distribution-level objective for generator post-training, complementing the conventional sample-le

APCReg: Anatomical-Prior-Guided Coarse-to-Fine CBCT--IOS Registration via Multi-View Projection and Reliability-Controlled Residual Correction

arXiv:2608.09993v1 Announce Type: cross Abstract: Registration between cone-beam computed tomography (CBCT) and intraoral scans (IOS) is essential for patient-specific surgical planning. However, disp

Beyond Forecasting: Recasting Volatility Control as a Routing Problem

arXiv:2608.10375v1 Announce Type: cross Abstract: Volatility control converts risk estimates into portfolio exposure, yet existing approaches often rely on a fixed volatility estimator or a pre-define

Big Tech's reliance on OpenAI and Anthropic for growth is systemically pervasive. Both companies are massively unprofitable — and much of wh…

Big Tech's reliance on OpenAI and Anthropic for growth is systemically pervasive. Both companies are massively unprofitable — and much of what they spend is Big Tech's own money, counted right back as

safetygary-marcus--x

BooST: Bridging Semantics and Motions for Efficient Skill Transfer

arXiv:2608.10600v1 Announce Type: cross Abstract: Skill abstraction---the process of learning reusable and temporally extended behaviors---has emerged as a key paradigm for improving sample efficiency

Model Releases

Boundary-Seeking Policy Gradient for Safe Reinforcement Learning

model-releasesarxiv-cs-lg

arXiv:2608.10204v1 Announce Type: new Abstract: Safe reinforcement learning maximizes reward subject to safety constraints. For Constrained Markov Decision Processes, the linear-programming view over

CARE: Confidence-Aware Reasoning for Reliable Medical VQA

arXiv:2608.10964v1 Announce Type: cross Abstract: Reinforcement Fine-Tuning (RFT) has enabled medical Multimodal Large Language Models (MLLMs) to produce Chain-of-Thought (CoT) reasoning for visual qu

Carefully Considering Culture: Analyzing LLM Alignment in Single- and Multi-Cultural Settings using Cultural Consensus Theory

arXiv:2608.09937v1 Announce Type: new Abstract: Recent work in NLP has probed large language models for their understanding of cultural norms across countries. However, this work typically considers d

ConRub-Med: Reinforcement Learning with Consensus Rubrics for Open-Ended Medical Question Answering

arXiv:2608.10996v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards has been especially effective in mathematics and coding, where answers can be checked automatically. Many

Convergence of Sign-based Random Reshuffling Algorithms for Nonconvex Optimization

arXiv:2310.15976v4 Announce Type: replace Abstract: signSGD is attractive in nonconvex optimization because it communicates sign-valued rather than full-precision gradients. Several standard analyses

Critic-Free Pretraining for Efficient Online Reinforcement Learning Fine-Tuning

arXiv:2608.10473v1 Announce Type: cross Abstract: Offline-to-online (O2O) reinforcement learning aims to leverage policies pretrained on static datasets while improving them through online interaction

Detecting an Effect Is Not Learning to Act on It: A Reward-SNR Floor for LLM Acquisition Agents

arXiv:2608.10441v1 Announce Type: cross Abstract: Many pipelines can pay a per-example cost to acquire an auxiliary, model-derived observation -- an LLM's structured reasoning, a slow oracle, an expen

DIMOS: Disentangling Instance-level Moving Object Segmentation

arXiv:2606.12826v2 Announce Type: replace-cross Abstract: Moving instance segmentation (MIS) attracts increasing attention due to its broad applications in traffic surveillance, autonomous driving, an

Do AI weather models miss extremes?

arXiv:2608.09972v1 Announce Type: cross Abstract: First-generation AI weather models are often reported to underperform at extremes, mostly in reanalysis-based evaluations of deterministic regression

Do Time-Series Forecasters Use the Right History: Recoverability, Recovery, and Functional Use of Temporal Delays

arXiv:2608.10433v1 Announce Type: new Abstract: Forecast accuracy does not tell us which past inputs produced a prediction. We separate three questions for time-series models with known delay structur

Dual-Loop Self-Evolution via Verifiable Emotion Feedback for Multi-Turn Empathetic Dialogue

arXiv:2608.10626v1 Announce Type: new Abstract: Large language models have demonstrated conversational capabilities, yet empathetic competence remains challenging. Empathetic support is inherently mul

Dual Space Preconditioning for Gradient Descent in the Overparameterized Regime

arXiv:2603.10485v3 Announce Type: replace-cross Abstract: In this work, we study the convergence properties of the Dual Space Preconditioned Gradient Descent, encompassing optimizers such as Normalize

Efficient Hypergradient Descent for Inverse Reinforcement Learning

arXiv:2608.11052v1 Announce Type: new Abstract: Inverse reinforcement learning (IRL) aims to recover a reward function under which the resulting policy reproduces the behavior observed in expert demon

Enhancing Automated Essay Scoring With Three Techniques: Two-Stage Fine-Tuning, Score Alignment, and Self-Training

arXiv:2602.01747v2 Announce Type: replace Abstract: Automated Essay Scoring (AES) plays a crucial role in education by providing scalable and efficient assessment tools. However, in real-world setting

Evaluation-Conditioned Training: Teaching Models to Generalize to Stronger Oversight Regimes

arXiv:2608.10209v1 Announce Type: new Abstract: Feedback signals used to train Large Language Models (LLMs) are the primary driver of their behavior and our main lever for instilling alignment with hu