Search: “safety”

Probing the Geometry of Color Representations in Multilingual Language Models

arXiv:2610.18331v1 Announce Type: cross Abstract: Do multilingual language models learn a universal internal geometry for color, and is it perceptually grounded or primarily linguistic? We decompose c

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Process Rewards for Outcome-Guided Reasoning Steps in LLM Reinforcement Learning

DGX agent

arXiv:2604.02341v2 Announce Type: replace Abstract: Mathematical reasoning in large language models has improved substantially with reinforcement learning using verifiable rewards, where final answers

safetyarxiv-cs-lg

14 Oct 2026

Prompt Interference in LLM Post-Training: Why Pass@k Optimization Can Degrade Pass@1

arXiv:2602.21189v3 Announce Type: replace Abstract: Pass@k is a widely used performance metric for verifiable large language model tasks, including mathematical reasoning, code generation, and short-a

PrT-Flow: Probability-Tractable Flow Matching for Reinforcement Learning

arXiv:2610.17667v1 Announce Type: new Abstract: Diffusion models and Flow Matching can represent flexible distributions over robot actions, but many reinforcement learning algorithms also require the

PTGPO: Prefix Tree Group Policy Optimization

arXiv:2610.14737v1 Announce Type: new Abstract: Training language agents on long-horizon interactive tasks is challenging because sparse terminal rewards provide weak credit signals for intermediate a

[Re] Boosting the Visual Interpretability of CLIP via Adversarial Fine-Tuning

arXiv:2610.14832v1 Announce Type: new Abstract: This paper presents a reproducibility study of ``Boosting the Visual Interpretability of CLIP via Adversarial Fine-Tuning'', published at ICLR 2025, whi

ReAP: From Visual Failure Recognition to Action-Conditioned Risk Prediction for Runtime Failure Monitoring

arXiv:2610.16309v1 Announce Type: new Abstract: Despite rapid progress, generalist robot policies remain brittle during execution. Reliable deployment therefore requires runtime monitoring that provid

Recovering Low-Bit Failures in Vision-Language-Action Policies via Quantization-Conditioned Reinforcement Learning

arXiv:2610.15675v1 Announce Type: cross Abstract: Quantization lowers deployment costs of vision-language-action (VLA) policies but can compromise task completion. Across five VLA models and 4,050 rea

Red-Teaming of Vision-Language-Action Models via Quality Diversity Environment Generation

arXiv:2610.14415v1 Announce Type: new Abstract: A long-standing goal in robotics is to develop generalist robots that can perform tasks across a wide range of settings, flexibly specified with images

Redesigning Regularization for Effective Policy Smoothing

arXiv:2606.13169v2 Announce Type: replace Abstract: This paper proposes a novel regularization design to effectively smooth policy functions in reinforcement learning. While regularization that enhanc

REEL: Learning to Retrieve, Evaluate, and Evolve Visual Memory for Any Long-Horizon Multi-Shot Video Generation

arXiv:2610.16632v1 Announce Type: new Abstract: Memory-based long-video generation aims to combine user control with long-range visual consistency. We organize this process into three functional compo

REFLEX: A Learned On-Board Reflex for Robust Autonomous Collision Recovery of Soft Quadrotors

arXiv:2610.15867v1 Announce Type: new Abstract: Collision-resilient airframes allow quadrotors to survive impacts, but do not ensure autonomous post-collision recovery. Innovation gating in stock onbo

Registration is a Representation Problem: Solving Whole-Body PET/CT Registration

arXiv:2610.18804v1 Announce Type: new Abstract: Raw image intensities are a poor surrogate for similarity across images in correspondence problems. Further, while auxiliary segmentation labels can pro

Reinforcement Learning Based Heterogeneous Task Allocation Under Spatio-Temporal Environmental Uncertainty for Multi-Robot Teams

arXiv:2610.15237v1 Announce Type: new Abstract: Autonomous Surface Vehicles (ASVs) have been increasingly deployed to perform different kinds of maritime tasks in recent years, which motivates the nee

Reinforcement Learning for Large-Scale Lifelong Multi-Agent Path Finding with Shared Spatial Fields

arXiv:2610.18186v1 Announce Type: new Abstract: Large-scale applications of lifelong multi-agent path finding (LMAPF) call for computationally inexpensive movement decisions under changing traffic con

Reinforcement Learning from Real-Robot Interaction without High-Quality State Estimates

arXiv:2610.14042v1 Announce Type: new Abstract: A longstanding goal for robot reinforcement learning (RL) is that robots learn from their own physical experience, as humans and animals do. A key obsta

Reinforcement Learning with Hawkes-Structured Networks for Market Making

arXiv:2610.16524v1 Announce Type: new Abstract: Market making under realistic order-flow dynamics is challenging due to the clustered, self-exciting nature of order arrivals, which are naturally model

Reliable Scale-Free Bandits via Empirical Likelihood

arXiv:2610.18055v1 Announce Type: cross Abstract: We study stochastic bandits without knowledge of reward scales or tail parameters. Almost-sure regret guarantees describe eventual learning but do not

Representation Geometry and Inductive Biases in Quantum Classifiers across Classical Visual Benchmarks

arXiv:2610.17722v1 Announce Type: cross Abstract: Quantum Machine Learning (QML) has a theoretical advantage over classical algorithms for specific synthetic data that is computationally hard. However

RePrior: Measurement-Reconciled Prior Correction for Sparse-View CBCT Reconstruction

arXiv:2610.14102v1 Announce Type: cross Abstract: Sparse-view cone-beam computed tomography (CBCT) is severely under-constrained by limited X-ray projections. Learned reconstruction can provide comple

Residual Compliant Policy: Learning Compliant Policies from Few Demonstrations

arXiv:2610.14035v1 Announce Type: new Abstract: Contact-rich manipulation requires a robot to be precise in free space to establish the intended contacts, and compliant once in contact to avoid excess

RESTA: Time-Scaling Augmentation for Data-Efficient Robot Imitation Learning

arXiv:2610.18163v1 Announce Type: new Abstract: Robot manipulation policies trained through imitation learning commonly rely on costly demonstrations recorded in one task direction and at one temporal

Rethinking Activation Functions: Reducing Bias Can Improve Learning

arXiv:2610.14500v1 Announce Type: new Abstract: Many widely used activation functions, including ReLU, GELU, Swish, and Sigmoid, produce predominantly positive outputs, resulting in a nonzero post-act

Rethinking Human Intent-to-CAD: Parametric CAD Model Generation via Cooperative Multi-Task Alignment and Spatial-Aware Reinforcement Learning

arXiv:2610.15681v1 Announce Type: new Abstract: Parametric Computer-Aided-Design (CAD) modeling from human intent remains challenging, particularly during the conceptual design stage, where design goa

Rethinking Projector Training in Multimodal LLMs

arXiv:2610.18598v1 Announce Type: new Abstract: Multimodal LLMs are typically trained by first aligning a vision encoder with a language model via projector training on large image-caption datasets, f

Revisiting Natural Language Prompts from Large Language Models for Open-Vocabulary Video Recognition

arXiv:2610.16666v1 Announce Type: new Abstract: Vision-language models (VLMs) like CLIP have shown great success in image tasks, however, applying CLIP to open-vocabulary video recognition poses chall

Reward-Aligned Distillation

arXiv:2610.14572v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards improves the reasoning capabilities of large language models, but remains computationally expensive and s

Risk-Averse Inference-Time Diffusion Model Alignment with Finite-Sample Guarantees

arXiv:2610.17579v1 Announce Type: cross Abstract: Inference-time alignment steers diffusion models toward downstream objectives without costly fine-tuning, but existing methods typically optimize expe

RoboFQE: Rollout-Free Evaluation of Visual Robot Policies from Offline Trajectories

arXiv:2610.18699v1 Announce Type: new Abstract: Evaluating robot policies before deployment typically requires producing new trajectories through real-world execution, simulators, or learned world mod

SARATHI: Decoupled Data-Free Probing and Adaptive Reconstruction for Structured LLM Pruning

arXiv:2610.15514v1 Announce Type: new Abstract: Structured pruning reduces the inference cost of large language models (LLMs) by permanently removing neurons and weight rows, yielding dense, hardware-

Scalable Multi-Human Cooperative Manipulation via Commander-Executor Coordination

arXiv:2610.14911v1 Announce Type: cross Abstract: Multi-human cooperative object manipulation is a highly challenging topic in humanoid control. While existing methods can handle simple collaborative

Scaling Latent Motor Adaptation to Coordinated Bimanual Manipulation

arXiv:2610.14833v1 Announce Type: new Abstract: Robust manipulation under uncertain object dynamics and geometry remains a core challenge in real-world robotics. Policies trained under fixed assumptio

ScenePlacer: Reformulating Scene-Level Object Pose Estimation as Visibility Field Learning

arXiv:2610.15734v1 Announce Type: new Abstract: We present ScenePlacer, a framework for scene-level pose estimation of objects generated by state-of-the-art native shape generation models, placing gen

Selective Successor Eligibility Traces for RLVR

arXiv:2605.05965v2 Announce Type: replace Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has become a key approach for improving the reasoning abilities of large language models. Howe

Sequent-Prover: Training Agents for Formal, Checkable SMT-based Reasoning

arXiv:2610.15258v1 Announce Type: cross Abstract: We introduce Sequent-Prover, a family of verifier-using language-model agents, and the post-training framework that produces them. Agents reason in an

SHOES: Sim-to-real Hardware-adaptation via On-policy Expansion of Skills for Legged Robots

arXiv:2610.16457v1 Announce Type: new Abstract: Humans expand their locomotion capability by wearing appropriate footwear, such as hiking boots for mountains or skates for ice. Footwear changes the gr

Silence and Noise: Resolving Suppression and Overvaluation in Multimodal Entity Alignment

arXiv:2610.16106v1 Announce Type: new Abstract: Multimodal entity alignment (MMEA) is fundamental to integrating heterogeneous knowledge graphs (KGs), yet existing paradigms remain critically vulnerab

Simpler and Stronger Baselines for All-in-One Image Restoration

arXiv:2512.10581v2 Announce Type: replace Abstract: Current methods for all-in-one image restoration increasingly rely on complex mechanisms, such as Mixture-of-Experts or prompt learning, to handle d

SkillScaffold: Scaffolding Agentic Reinforcement Learning with Policy-Adaptive Skill Credit Assignment and Selective Injection

arXiv:2610.18807v1 Announce Type: new Abstract: Agentic reinforcement learning (RL) enables large language models (LLMs) to learn from multi-turn interactions with verifiable outcomes. Textual skills

SlotAVS: Object-Centric Audio-Visual Segmentation via Bi-Modal Slot Attention

arXiv:2610.15764v1 Announce Type: new Abstract: Audio-Visual Segmentation (AVS) requires predicting pixel-wise masks of sounding objects by grounding acoustic signals in visual space. In multi-source

Small area learning

arXiv:2610.15953v1 Announce Type: cross Abstract: Small area estimation (SAE) traditionally borrows strength across domains through a prespecified working model. We introduce small area learning (SAL)

Smoothness Ranks Skill, Not Success: An Audit of Published Trajectory-Smoothness Curation Metrics Under Episode-Length Controls

arXiv:2610.17427v1 Announce Type: new Abstract: Trajectory smoothness has been proposed as a cheap, training-free signal for curating imitation-learning demonstrations, most recently by rinse, which i

Social Priors: Descriptions of AI Users Shape How Language Models Trust and Cooperate

arXiv:2610.16946v1 Announce Type: new Abstract: Descriptions of how people use AI are becoming training data for the next generation of models. We ask whether what a model reads about its users change

SPARCL: Spatially Adaptive Regional Coverage for Active Learning in Chest X-ray Analysis

arXiv:2610.17808v1 Announce Type: new Abstract: Pixel-level annotation for chest X-ray (CXR) lesion segmentation is expensive, making active learning (AL) essential. However, existing AL strategies fa

Sparse-TD-JEPA: Improving Zero-Shot Reinforcement Learning With Sparse Latent Task Representations

arXiv:2610.14400v1 Announce Type: new Abstract: Successor-feature-based zero-shot reinforcement learning methods pre-train a task encoder and successor-feature predictor on offline, reward-free data,

Splicing Hints into Reasoning Traces: Learning from Demonstrator Failures

arXiv:2610.13996v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards improves mathematical reasoning when a final answer can be checked, but a correct answer does not certify

SPROUT: Scaling Agentic Reinforcement Learning for Long-Horizon Software Engineering

arXiv:2610.14405v1 Announce Type: cross Abstract: Reinforcement learning (RL) has driven substantial progress in coding agents by relying on verified successful trajectories as learning signals. Howev

Start Right, Arrive Right: Asynchronous Execution via Initial Noise Selection

arXiv:2606.19774v2 Announce Type: replace Abstract: Action chunking enables robot policies to produce temporally coherent behavior, but generating multi-step action sequences with flow-based policies