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

Jan 28

Submitted by

jasonrqh

AgentDoG: A Diagnostic Guardrail Framework for AI Agent Safety and Security

AI45Research

Submitted by

hitsmy

AdaReasoner: Dynamic Tool Orchestration for Iterative Visual Reasoning

Fudan-University

Fudan University

Submitted by

zkcys001

A Pragmatic VLA Foundation Model

robbyant

Submitted by

mrdbourke

Youtu-VL: Unleashing Visual Potential via Unified Vision-Language Supervision

tencent

Submitted by

manchery

Visual Generation Unlocks Human-Like Reasoning through Multimodal World Models

ByteDance-Seed

Submitted by

xi-j

AVMeme Exam: A Multimodal Multilingual Multicultural Benchmark for LLMs' Contextual and Cultural Knowledge and Thinking

columbia

Columbia University

Submitted by

kpzhang996

World Craft: Agentic Framework to Create Visualizable Worlds via Text

ShandaAI

Shanda AI Research Tokyo

Submitted by

laiwei-seed

Post-LayerNorm Is Back: Stable, ExpressivE, and Deep

ByteDance-Seed

3

Submitted by

sthui

Towards Pixel-Level VLM Perception via Simple Points Prediction

moonshotai

Submitted by

lincharliesun

FABLE: Forest-Based Adaptive Bi-Path LLM-Enhanced Retrieval for Multi-Document Reasoning

qihoo360

北京奇虎科技有限公司

3

Submitted by

lincharliesun

TriPlay-RL: Tri-Role Self-Play Reinforcement Learning for LLM Safety Alignment

qihoo360

北京奇虎科技有限公司

3

Submitted by

phyang

HalluCitation Matters: Revealing the Impact of Hallucinated References with 300 Hallucinated Papers in ACL Conferences

·
3 authors

Submitted by

quyanh

Selective Steering: Norm-Preserving Control Through Discriminative Layer Selection

knoveleng

Knovel Engineering

Submitted by

ufotalent

Revisiting Parameter Server in LLM Post-Training

sail

Submitted by

improbableaimit

Self-Distillation Enables Continual Learning

MIT

Massachusetts Institute of Technology

Submitted by

dginf

HyperAlign: Hypernetwork for Efficient Test-Time Alignment of Diffusion Models

·
3 authors

2

Submitted by

weiber2002

EvolVE: Evolutionary Search for LLM-based Verilog Generation and Optimization

nationaltaiwan

Submitted by

martheballon

Benchmarks Saturate When The Model Gets Smarter Than The Judge

·
4 authors

3

Submitted by

ethanning

GPCR-Filter: a deep learning framework for efficient and precise GPCR modulator discovery

CarnegieMellonU

Carnegie Mellon University

4

Submitted by

manthanhp

DeFM: Learning Foundation Representations from Depth for Robotics

leggedrobotics

Robotic Systems Lab - ETH Zürich

2

Submitted by

ProKil

CooperBench: Why Coding Agents Cannot be Your Teammates Yet

stanfordnlp