Kiều Sơn Tùng

codemaivanngu's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

upvoted a paper about 20 hours ago

How Far Are We from Removing the Visual Encoder? Scaling Laws for Encoder-Free Multimodal Pretraining

Paper • 2609.35457 • Published 4 days ago • 65

upvoted 3 papers 2 days ago

Learning Native Reflection in Unified Models with Interleaved Reinforcement Learning

Paper • 2609.35767 • Published 4 days ago • 47

Just MLPs: Efficient Visual State Reconstruction for Multimodal Language Models

Paper • 2609.34972 • Published 4 days ago • 28

Rethinking Training-Inference Mismatch in LLM Reinforcement Learning: Where It Arises and How to Correct It

Paper • 2609.32444 • Published 6 days ago • 26

upvoted a paper 7 days ago

PACT: From Credit Assignment to Critic Alignment

Paper • 2609.26355 • Published 10 days ago • 29

upvoted a paper 20 days ago

SenseNova-U1.5: Towards Native Unified Visual Intelligence

Paper • 2609.11929 • Published 22 days ago • 277

upvoted a paper 21 days ago

SpatialBlock: Enhancing Spatial Intelligence in LVLMs via Synthetic Block-Stacking Problem

Paper • 2609.07064 • Published 25 days ago • 147

upvoted a paper 23 days ago

Causal Foundation Models

Paper • 2609.03003 • Published about 1 month ago • 30

upvoted a paper 24 days ago

Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference

Paper • 2609.05275 • Published 28 days ago • 26

upvoted a paper about 1 month ago

TTPO: Test-Time Policy Optimization

Paper • 2608.27448 • Published Aug 27 • 84

upvoted a collection about 1 month ago

Granite 4.2 Language Models

Collection

Efficient reasoning and thinking language models for multilingual generation, coding, and AI assistant workflows. • 24 items • Updated 17 days ago • 43

upvoted 4 papers about 1 month ago

Learn What's Left, Not What's Mastered: Saturation Aware Advantage Reweighting for Multi-Reward Policy Optimization

Paper • 2608.16072 • Published Aug 17 • 52

DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data

Paper • 2608.13517 • Published Aug 13 • 33

Improving the matrix multiplication exponent with modern optimization and AlphaEvolve

Paper • 2608.16884 • Published Aug 17 • 20

Modular Cognitive Architecture Emerges in Large Language Models

Paper • 2608.13567 • Published Jun 27 • 16

upvoted a paper about 2 months ago

Stealing Reasoning Traces from Proprietary LLM APIs

Paper • 2608.09867 • Published Aug 10 • 122

upvoted 4 papers 2 months ago

Mage-VL: An Efficient Codec-Native Streaming Multimodal Foundation Model

Paper • 2607.24904 • Published Jul 27 • 37

Chimera: Designing and Chinchilla-Scaling Hybrid Visual Diffusion Transformers

Paper • 2607.28611 • Published Jul 30 • 23

Multi-Head Attention Residuals

Paper • 2607.27230 • Published Jul 22 • 13

Molt: A Scalable PyTorch-Native Training Framework for Agentic Reinforcement Learning

Paper • 2607.21653 • Published Jul 22 • 34