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

Feb 23

Submitted by

floyed

VESPO: Variational Sequence-Level Soft Policy Optimization for Stable Off-Policy LLM Training

rednote-hilab

Submitted by

Yikunb

Does Your Reasoning Model Implicitly Know When to Stop Thinking?

ByteDance

5

Submitted by

taesiri

Generated Reality: Human-centric World Simulation using Interactive Video Generation with Hand and Camera Control

·
6 authors

Submitted by

hba123

Decoding as Optimisation on the Probability Simplex: From Top-K to Top-P (Nucleus) to Best-of-K Samplers

·
4 authors

Submitted by

hilamanor

Spanning the Visual Analogy Space with a Weight Basis of LoRAs

nvidia

Submitted by

taesiri

EgoPush: Learning End-to-End Egocentric Multi-Object Rearrangement for Mobile Robots

·
7 authors

Submitted by

taesiri

SARAH: Spatially Aware Real-time Agentic Humans

·
5 authors

Submitted by

nielsr

VidEoMT: Your ViT is Secretly Also a Video Segmentation Model

tue-mps

Mobile Perception Systems Lab

Submitted by

aidar-myrzakhan

Sink-Aware Pruning for Diffusion Language Models

MBZUAI

Mohamed Bin Zayed University of Artificial Intelligence

Submitted by

skylenage

DeepVision-103K: A Visually Diverse, Broad-Coverage, and Verifiable Mathematical Dataset for Multimodal Reasoning

·
8 authors

Submitted by

rainnekoneko

Avey-B

avey-ai

Submitted by

taesiri

Learning Smooth Time-Varying Linear Policies with an Action Jacobian Penalty

·
3 authors

Submitted by

beopst

Selective Training for Large Vision Language Models via Visual Information Gain

SeoulTech

Seoul National University of Science and Technology

Submitted by

minxin-zhang

Adam Improves Muon: Adaptive Moment Estimation with Orthogonalized Momentum

UCLA

University of California, Los Angeles

2

Submitted by

takyoung

ReIn: Conversational Error Recovery with Reasoning Inception

illinois

University of Illinois at Urbana-Champaign

Submitted by

Czardas

Whom to Query for What: Adaptive Group Elicitation via Multi-Turn LLM Interactions

ZD Lab @ UNC-Chapel Hill

Submitted by

Czardas

Rubrics as an Attack Surface: Stealthy Preference Drift in LLM Judges

ZD Lab @ UNC-Chapel Hill

Submitted by

Luo-Yihang

4RC: 4D Reconstruction via Conditional Querying Anytime and Anywhere

·
5 authors