new

Get trending papers in your email inbox once a day!

Get trending papers in your email inbox!

Daily Papers

byAK and the research community

Jun 10

Submitted by

qian43

ABot-Earth 0.5: Generative 3D Earth Model

acvlab

Alibaba AMAP CV Lab

Submitted by

liangtm

Kwai Keye-VL-2.0 Technical Report

KwaiKeye

Submitted by

xuc865

Role-Agent: Bootstrapping LLM Agents via Dual-Role Evolution

·
7 authors

Submitted by

wenbopan

Retrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory Rollouts

MicrosoftResearch

Microsoft Research

Submitted by

QQqqrabit

SearchSwarm: Towards Delegation Intelligence in Agentic LLMs for Long-Horizon Deep Research

SearchSwarm

Submitted by

YZCS

MemDreamer: Decoupling Perception and Reasoning for Long Video Understanding via Hierarchical Graph Memory and Agentic Retrieval Mechanism

inclusionAI

Submitted by

P2333

Flow-DPPO: Divergence Proximal Policy Optimization for Flow Matching Models

Tencent-Hunyuan-Multimodal-RL

Tencent-Hunyuan-Multimodal-RL

Submitted by

taesiri

SCAIL-2: Unifying Controlled Character Animation with End-to-end In-Context Conditioning

·
4 authors

Submitted by

YJ-142150

Lip Forcing: Few-Step Autoregressive Diffusion for Real-time Lip Synchronization

kaist-ai

Submitted by

hzeroyuke

WorldOlympiad: Can Your World Model Survive a Triathlon?

Alibaba-DAMO-Academy

Submitted by

zhouxiangxin

Rethinking the Divergence Regularization in LLM RL

Tencent-Hunyuan-Multimodal-RL

Tencent-Hunyuan-Multimodal-RL

Submitted by

Daniel0724

ARM: An AutoRegressive Large Multimodal Model with Unified Discrete Representations

·
19 authors

Submitted by

ShilongLiu

EEVEE: Towards Test-time Prompt Learning in the Real World for Self-Improving Agents

princeton-ai

Princeton AI Lab

Submitted by

taesiri

Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields

ByteDance-Seed

Submitted by

zz1358m

One Token per Multimodal Evidence: Latent Memory for Resource-Constrained QA

NationalUniversityofSingapore

National University of Singapore

Submitted by

xinyu04

Attention Amnesia in Hybrid LLMs: When CoT Fine-Tuning Breaks Long-Range Recall, and How to Fix It

LARK-Lab

LARK Lab@HKUST (GZ)

Submitted by

elephantmipt

Interpreting and Steering a Text-to-Speech Language Model with Sparse Autoencoders

t-tech

Submitted by

plusn

Online Skill Learning for Web Agents via State-Grounded Dynamic Retrieval

UGA-AI

University of Georgia

Submitted by

nnnyt

SkillHarm: Lifecycle-Aware Skill-Based Attacks via Automated Construction

osunlp

Submitted by

yangcole

How Does Reasoning Flow? Tracing Attention-Induced Information Flow for Targeted RL in LLMs

alibabagroup

Submitted by

giannor

BrainSurgery: Reproducible and Reliable Declarative Weight Manipulations for Model Editing and Upcycling

SDU-Denmark

University of Southern Denmark (SDU)

Submitted by

X1AOX1A

Bridging the Agent-World Gap: Text World Models for LLM-based Agents

SUSTech-NLP

NLP Group in SUSTech

Submitted by

giannor

PsychoSafe: Eliciting Psychologically-Informed Refusals in Large Language Models

SDU-Denmark

University of Southern Denmark (SDU)

Submitted by

Karl28

Struct-Searcher: Agentic Structural Thinking Advances Multimodal Deep Information Seeking

CUHK-CSE

The Chinese University of Hong Kong

Submitted by

upyzwup

UniPET: a universal network for high-quality PET image denoising across varied dose reduction factors

·
7 authors

Submitted by

upyzwup

U-TTT: Towards Generalizable PET Image Denoising via Test-Time Training

·
7 authors

Submitted by

taesiri

MilliVid: Hierarchical Latents for Long-Range Consistency in Video Generation

·
7 authors

Submitted by

yyy127

Emergent Misalignment Can Be Induced by Sycophancy and Reversed via Alignment Gating

Submitted by

Albus-Chen

What Should Agents Say? Action-state Communication for Efficient Multi-Agent Systems

iNLP-Lab

iNLP Lab @ SUTD

Submitted by

gangweix

Next Forcing: Causal World Modeling with Multi-Chunk Prediction

·
7 authors

Submitted by

taesiri

Dynamic Linear Attention

ByteDance-Seed

Submitted by

Jinyang23

Late-Layer Fusion is Enough: Dual-Path Vision Token Routing for Multimodal Large Language Models under Visual Saturation

·
2 authors

Submitted by

Erosinho

IR3DE: A Linear Router for Large Language Models

Gensyn

Submitted by

semihkara

The Role of Feedback Alignment in Self-Distillation

Gensyn

Submitted by

taesiri

Test-Time Gradient Guidance of Flow Policies in Reinforcement Learning

·
7 authors

Submitted by

skydddoogg

Do Coding Agents Deceive Us? Detecting and Preventing Cheating via Capped Evaluation with Randomized Tests

ishidalab

Submitted by

UVSKKR

When the Chain of Thought Knows Better: Failure Modes in Multi-Turn Reasoning Models

·
3 authors

Submitted by

gist-sparse-attention

Decentralized Multi-Agent Systems with Shared Context

StanfordUniversity

Stanford University

Submitted by

Jerry999

PaperMentor: A Human-Centered Multi-Agent Writing Tutor for AI Research Papers on Overleaf

Jinesis

Submitted by

EnyiJiang

When Behavioral Safety Evaluation Fails: A Representation-Level Perspective

StanfordUniversityy

Stanford University

Submitted by

Simase

FadeMem: Distance-Aware Memory Consolidation for Autoregressive Video Diffusion

·
5 authors

Submitted by

Sajib-006

BenSyc: Benchmarking Conversational Sycophancy and Human Alignment in LLMs for Bengali Contexts

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

marvinsxtr

In-Context Multiple Instance Learning

·
5 authors