Search: “models”

Learning Agentic Policy from Action Guidance

arXiv:2605.12004v1 Announce Type: new Abstract: Agentic reinforcement learning (RL) for Large Language Models (LLMs) critically depends on the exploration capability of the base policy, as training si

Leveraging RAG for Training-Free Alignment of LLMs

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

arXiv:2605.11217v1 Announce Type: new Abstract: Large language model (LLM) alignment algorithms typically consist of post-training over preference pairs. While such algorithms are widely used to enabl

safetyarxiv-cs-lg

13 May 2026

Localization Boosting for Growth Markets: Mitigating Cross-Locale Behavioral Bias in Learning-to-Rank

arXiv:2605.11272v1 Announce Type: new Abstract: Adobe Express is expanding internationally, but the US has a disproportionately large content supply and interaction volume. Learning-to-rank (LTR) mode

local-aiarxiv-cs-lg

Logit-Attention Divergence: Mitigating Position Bias in Multi-Image Retrieval via Attention-Guided Calibration

arXiv:2605.11591v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) have shown strong performance in multi-image cross-modal retrieval, yet suffer from severe position bias, where

safetyarxiv-cs-cv

Maximin Robust Bayesian Experimental Design

arXiv:2603.14094v2 Announce Type: replace-cross Abstract: We address the brittleness of Bayesian experimental design under model misspecification by formulating the problem as a max--min game between

MCPShield: Content-Aware Attack Detection for LLM Agent Tool-Call Traffic

arXiv:2605.11053v1 Announce Type: cross Abstract: The Model Context Protocol (MCP) has become a widely adopted interface for LLM agents to invoke external tools, yet learned monitoring of MCP tool-cal

agentsarxiv-cs-lg

Mitigating Action-Relation Hallucinations in LVLMs via Relation-aware Visual Enhancement

arXiv:2605.11808v1 Announce Type: new Abstract: Large Vision-Language Models (LVLMs) have achieved remarkable performance on diverse vision-language tasks. However, LVLMs still suffer from hallucinati

local-aiarxiv-cs-cv

OGLS-SD: On-Policy Self-Distillation with Outcome-Guided Logit Steering for LLM Reasoning

arXiv:2605.12400v1 Announce Type: new Abstract: We study {on-policy self-distillation} (OPSD), where a language model improves its reasoning ability by distilling privileged teacher distributions alon

Applications

Online Learning-to-Defer with Varying Experts

arXiv:2605.12340v1 Announce Type: cross Abstract: Learning-to-Defer (L2D) methods route each query either to a predictive model or to external experts. While existing work studies this problem in batc

applicationsarxiv-cs-lg

OUI as a Structural Observable: Towards an Activation-Centric View of Neural Network Training

arXiv:2605.11570v1 Announce Type: new Abstract: Activation functions are what make deep networks expressive: without them, the model collapses to a linear map. Yet we still evaluate training mostly fr

Physics Aware Neural Networks: Denoising for Magnetic Navigation

arXiv:2602.13690v2 Announce Type: replace Abstract: Magnetic-anomaly navigation, leveraging small-scale variations in the Earth's magnetic field, is a promising alternative when GPS is unavailable or

Pion: A Spectrum-Preserving Optimizer via Orthogonal Equivalence Transformation

arXiv:2605.12492v1 Announce Type: new Abstract: We introduce Pion, a spectrum-preserving optimizer for large language model (LLM) training based on orthogonal equivalence transformation. Unlike additi

PIVOT: Bridging Planning and Execution in LLM Agents via Trajectory Refinement

arXiv:2605.11225v1 Announce Type: cross Abstract: Large language model (LLM)-based agents frequently generate seemingly coherent plans that fail upon execution due to infeasible actions, constraint vi

agentsarxiv-cs-lg

Predicting Disagreement with Human Raters in LLM-as-a-Judge Difficulty Assessment without Using Generation-Time Probability Signals

arXiv:2605.12422v1 Announce Type: new Abstract: Automatic generation of educational materials using large language models (LLMs) is becoming increasingly common, but assigning difficulty levels to suc

researcharxiv-cs-cl

Predictive Maps of Multi-Agent Reasoning: A Successor-Representation Spectrum for LLM Communication Topologies

arXiv:2605.11453v1 Announce Type: cross Abstract: Practitioners deploying multi-agent large language model (LLM) systems must currently choose between communication topologies such as chain, star, mes

Principle-Guided Supervision for Interpretable Uncertainty in Medical Image Segmentation

arXiv:2605.10984v1 Announce Type: new Abstract: Uncertainty quantification complements model predictions by characterizing their reliability, which is essential for high-stakes decision making such as

RACC: Representation-Aware Coverage Criteria for LLM Safety Testing

arXiv:2602.02280v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) face severe safety risks from jailbreak attacks, yet current safety testing largely relies on static datasets and

Robust Biomedical Publication Type and Study Design Classification with Knowledge-Guided Perturbations

arXiv:2605.11502v1 Announce Type: new Abstract: Accurately and consistently indexing biomedical literature by publication type and study design is essential for supporting evidence synthesis and knowl

researcharxiv-cs-cl

SI-Diff: A Framework for Learning Search and High-Precision Insertion with a Force-Domain Diffusion Policy

arXiv:2605.12247v1 Announce Type: new Abstract: Contact-rich assembly is fundamental in robotics but poses significant challenges due to uncertainties in relative poses, such as misalignments and smal

safetyarxiv-cs-ro

SkillGraph: Skill-Augmented Reinforcement Learning for Agents via Evolving Skill Graphs

arXiv:2605.12039v1 Announce Type: new Abstract: Skill libraries enable large language model agents to reuse experience from past interactions, but most existing libraries store skills as isolated entr

Sobolev Regularized MMD Gradient Flow

arXiv:2605.11884v1 Announce Type: new Abstract: We propose Sobolev-regularized Maximum Mean Discrepancy (SrMMD) gradient flow, a regularized variant of maximum mean discrepancy (MMD) gradient flow bas

SRG: Score-based Relaxation-guided Generation for Mixed Integer Linear Programming

arXiv:2603.24033v2 Announce Type: replace Abstract: We propose Score-based Relaxation-guided Generation (SRG), a generative framework based on an approximate formulation of relaxation-guided stochasti

Tackling Fake Forgetting through Uncertainty Quantification

arXiv:2501.19403v3 Announce Type: replace Abstract: Machine unlearning seeks to remove the influence of specified data from a trained model. While the unlearning accuracy provides a widely used metric

Taming Extreme Tokens: Covariance-Aware GRPO with Gaussian-Kernel Advantage Reweighting

arXiv:2605.11538v1 Announce Type: new Abstract: Group Relative Policy Optimization (GRPO) has emerged as a promising approach for improving the reasoning capabilities of large language models. However

Taming Score-Based Denoisers in ADMM: A Convergent Plug-and-Play Framework

arXiv:2603.10281v3 Announce Type: replace-cross Abstract: While score-based generative models have emerged as powerful priors for solving inverse problems, directly integrating them into optimization

The Algorithmic Caricature: Auditing LLM-Generated Political Discourse Across Crisis Events

arXiv:2605.12452v1 Announce Type: new Abstract: Large Language Models (LLMs) can generate fluent political text at scale, raising concerns about synthetic discourse during crises and social conflict.

researcharxiv-cs-cl

The tractability landscape of diffusion alignment: regularization, rewards, and computational primitives

arXiv:2605.11361v1 Announce Type: new Abstract: Inference-time reward alignment asks how to turn a pre-trained diffusion model with base law p into a sampler that favors a reward r while remaining clo

Toxicity Detection Should Measure Contextual Harm, Not Text-Intrinsic Badness

arXiv:2503.16072v4 Announce Type: replace-cross Abstract: Toxicity detection has become core safety infrastructure for online moderation, dataset filtering, and deployed language-model systems. Yet mo

Tutorials

UniCustom: Unified Visual Conditioning for Multi-Reference Image Generation

arXiv:2605.12088v1 Announce Type: new Abstract: Multi-reference image generation aims to synthesize images from textual instructions while faithfully preserving subject identities from multiple refere

tutorialsarxiv-cs-cv

UniFixer: A Universal Reference-Guided Fixer for Diffusion-Based View Synthesis

arXiv:2605.12169v1 Announce Type: new Abstract: With the recent surge of generative models, diffusion-based approaches have become mainstream for view synthesis tasks, either in an explicit depth-warp

safetyarxiv-cs-cv

Unlearning with Asymmetric Sources: Improved Unlearning-Utility Trade-off with Public Data

arXiv:2605.11170v1 Announce Type: new Abstract: Noise-based certified machine unlearning currently faces a hard ceiling: the noise magnitude required to certify unlearning typically destroys model uti

Unlocking Compositional Generalization in Continual Few-Shot Learning

arXiv:2605.11710v1 Announce Type: cross Abstract: Object-centric representations promise a key property for few-shot learning: Rather than treating a scene as a single unit, a model can decompose it i

Applications

X-Imitator: Spatial-Aware Imitation Learning via Bidirectional Action-Pose Interaction

arXiv:2605.12162v1 Announce Type: new Abstract: Effectively handling the interplay between spatial perception and action generation remains a critical bottleneck in robotic manipulation. Existing meth

applicationsarxiv-cs-ro

Zero-shot Human Pose Estimation using Diffusion-based Inverse solvers

arXiv:2510.02043v2 Announce Type: replace Abstract: Pose estimation refers to tracking a human's full body posture, including their head, torso, arms, and legs. The problem is challenging in practical

Applications

A Comparative Study of Machine Learning and Deep Learning for Out-of-Distribution Detection

arXiv:2605.10181v1 Announce Type: cross Abstract: Out-of-distribution (OOD) detection is essential for building reliable AI systems, as models that produce outputs for invalid inputs cannot be trusted

applicationsarxiv-cs-ai

A Unified Pair-GRPO Family: From Implicit to Explicit Preference Constraints for Stable and General RL Alignment

arXiv:2605.06375v1 Announce Type: cross Abstract: Large language model (LLM) alignment via reinforcement learning from human preferences (RLHF) suffers from unstable policy updates, ambiguous gradient

local-aiarxiv-cs-ai

Accurate Trajectory Tracking with MPCC for Flapping-Wing MAVs

arXiv:2605.06042v2 Announce Type: replace Abstract: Flapping-wing micro aerial vehicles offer quieter and safer operation than rotary-wing drones, yet achieving precise autonomous control of bird-scal

agentsarxiv-cs-ro

Ace-Skill: Bootstrapping Multimodal Agents with Prioritized and Clustered Evolution

arXiv:2605.08887v1 Announce Type: new Abstract: Self-evolving agents present a promising path toward continual adaptation by distilling task interactions into reusable knowledge artifacts. In practice

agentsarxiv-cs-ai

AdaSwitch: Adaptive Switching between Small and Large Agents for Effective Cloud-Local Collaborative Learning

arXiv:2410.13181v2 Announce Type: replace Abstract: Recent advancements in large language models (LLMs) have been remarkable. Users face a choice between using cloud-based LLMs for generation quality

local-aiarxiv-cs-cl

Agentic AI Scientists Are Not Built For Autonomous Scientific Discovery

arXiv:2605.08956v1 Announce Type: new Abstract: A growing body of work pursues AI scientists capable of end-to-end autonomous scientific discovery. This position paper argues that although they alread

agentsarxiv-cs-ai

AgentSlimming: Towards Efficient and Cost-Aware Multi-Agent Systems

arXiv:2605.08813v1 Announce Type: new Abstract: Large Language Model-based Multi-Agent Systems (MAS) have demonstrated remarkable capabilities in complex tasks. However, manually designing optimal com

agentsarxiv-cs-lg

Alignment as Jurisprudence

arXiv:2605.08416v1 Announce Type: new Abstract: Jurisprudence, the study of how judges should properly decide cases, and alignment, the science of getting AI models to conform to human values, share a

safetyarxiv-cs-ai

An Empirical Analysis of Calibration and Selective Prediction in Multimodal Clinical Condition Classification

arXiv:2603.02719v2 Announce Type: replace Abstract: As artificial intelligence systems move toward clinical deployment, ensuring reliable prediction behavior is fundamental for safety-critical decisio

ArchRAG: Attributed Community-based Hierarchical Retrieval-Augmented Generation

arXiv:2502.09891v4 Announce Type: replace-cross Abstract: Retrieval-Augmented Generation (RAG) has proven effective in integrating external knowledge into large language models (LLMs) for solving ques

researcharxiv-cs-ai

Attention Sinks in Diffusion Transformers: A Causal Analysis

arXiv:2605.09313v1 Announce Type: new Abstract: Attention sinks -- tokens that receive disproportionate attention mass -- are assumed to be functionally important in autoregressive language models, bu

safetyarxiv-cs-cv

Attribution-based Explanations for Markov Decision Processes

arXiv:2605.09780v1 Announce Type: new Abstract: Attribution techniques explain the outcome of an AI model by assigning a numerical score to its inputs. So far, these techniques have mainly focused on

researcharxiv-cs-ai

Augmented Equivariant Mesh Networks for Anatomical Segmentation

arXiv:2605.08172v1 Announce Type: new Abstract: Anatomical mesh segmentation requires models that operate directly on irregular surface geometry while remaining robust to arbitrary patient pose and me

AxiomOcean: Forecasting the Three-Dimensional Structure of the Upper Ocean

arXiv:2605.10455v1 Announce Type: new Abstract: Short-term ocean forecast skill depends strongly on the three-dimensional ocean structure of the upper ocean, which governs stratification, subsurface h