wei

fengwei's activity

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

upvoted a collection 14 days ago

SmolLM3 pretraining datasets

Collection

datasets used in SmolLM3 pretraining • 14 items • Updated 23 days ago • 22

upvoted a paper 22 days ago

Should We Still Pretrain Encoders with Masked Language Modeling?

Paper • 2507.00994 • Published 30 days ago • 74

upvoted 5 papers about 1 month ago

Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning

Paper • 2506.01939 • Published Jun 2 • 176

Reflect, Retry, Reward: Self-Improving LLMs via Reinforcement Learning

Paper • 2505.24726 • Published May 30 • 267

MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention

Paper • 2506.13585 • Published Jun 16 • 260

Reinforcement Pre-Training

Paper • 2506.08007 • Published Jun 9 • 252

The Diffusion Duality

Paper • 2506.10892 • Published Jun 12 • 38

upvoted 2 papers about 2 months ago

Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models

Paper • 2506.05176 • Published Jun 5 • 67

MiniCPM4: Ultra-Efficient LLMs on End Devices

Paper • 2506.07900 • Published Jun 9 • 87

upvoted 2 papers 2 months ago

MiniMax-Speech: Intrinsic Zero-Shot Text-to-Speech with a Learnable Speaker Encoder

Paper • 2505.07916 • Published May 12 • 130

Qwen3 Technical Report

Paper • 2505.09388 • Published May 14 • 267

upvoted 5 papers 3 months ago

RM-R1: Reward Modeling as Reasoning

Paper • 2505.02387 • Published May 5 • 78

ZeroSearch: Incentivize the Search Capability of LLMs without Searching

Paper • 2505.04588 • Published May 7 • 66

ReasonIR: Training Retrievers for Reasoning Tasks

Paper • 2504.20595 • Published Apr 29 • 55

UniversalRAG: Retrieval-Augmented Generation over Multiple Corpora with Diverse Modalities and Granularities

Paper • 2504.20734 • Published Apr 29 • 63

CLIMB: CLustering-based Iterative Data Mixture Bootstrapping for Language Model Pre-training

Paper • 2504.13161 • Published Apr 17 • 92

upvoted a paper 4 months ago

Kimi-VL Technical Report

Paper • 2504.07491 • Published Apr 10 • 133

upvoted a collection 4 months ago

Kimi-VL-A3B

Collection

Moonshot's efficient MoE VLMs, exceptional on agent, long-context, and thinking • 7 items • Updated about 1 month ago • 71

upvoted 2 papers 4 months ago

ZClip: Adaptive Spike Mitigation for LLM Pre-Training

Paper • 2504.02507 • Published Apr 3 • 90

Advances and Challenges in Foundation Agents: From Brain-Inspired Intelligence to Evolutionary, Collaborative, and Safe Systems

Paper • 2504.01990 • Published Mar 31 • 300