Peter Szemraj PRO

pszemraj's activity

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

upvoted a paper about 1 hour ago

A Survey on Latent Reasoning

Paper • 2507.06203 • Published 1 day ago • 63

upvoted an article about 2 hours ago

view article

Article

SmolLM3: smol, multilingual, long-context reasoner

By and 22 others •

2 days ago

• 412

upvoted a collection about 15 hours ago

T5Gemma

Collection

32 items • Updated about 15 hours ago • 27

upvoted an article about 15 hours ago

view article

Article

Reachy Mini - The Open-Source Robot for Today's and Tomorrow's AI Builders

By and 1 other •

1 day ago

• 331

upvoted 2 papers 1 day ago

Easy Dataset: A Unified and Extensible Framework for Synthesizing LLM Fine-Tuning Data from Unstructured Documents

Paper • 2507.04009 • Published 5 days ago • 19

Should We Still Pretrain Encoders with Masked Language Modeling?

Paper • 2507.00994 • Published 9 days ago • 71

upvoted a paper 3 days ago

Energy-Based Transformers are Scalable Learners and Thinkers

Paper • 2507.02092 • Published 7 days ago • 43

upvoted a paper 5 days ago

Self-Correction Bench: Revealing and Addressing the Self-Correction Blind Spot in LLMs

Paper • 2507.02778 • Published 7 days ago • 9

upvoted an article 8 days ago

view article

Article

Training and Finetuning Sparse Embedding Models with Sentence Transformers v5

By and 1 other •

9 days ago

• 85

upvoted 3 papers 12 days ago

Is There a Case for Conversation Optimized Tokenizers in Large Language Models?

Paper • 2506.18674 • Published 17 days ago • 8

Outlier-Safe Pre-Training for Robust 4-Bit Quantization of Large Language Models

Paper • 2506.19697 • Published 16 days ago • 44

Where to find Grokking in LLM Pretraining? Monitor Memorization-to-Generalization without Test

Paper • 2506.21551 • Published 14 days ago • 27

upvoted a collection 13 days ago

Gemma 3n

Collection

4 items • Updated about 15 hours ago • 163

upvoted a collection 17 days ago

Gemma 3 QAT

Collection

Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory • 15 items • Updated about 15 hours ago • 203

upvoted 2 papers 22 days ago

Scientists' First Exam: Probing Cognitive Abilities of MLLM via Perception, Understanding, and Reasoning

Paper • 2506.10521 • Published 28 days ago • 70

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

Paper • 2506.13585 • Published 24 days ago • 252

upvoted a paper 26 days ago

Domain2Vec: Vectorizing Datasets to Find the Optimal Data Mixture without Training

Paper • 2506.10952 • Published 28 days ago • 23

upvoted 2 papers 29 days ago

Through the Valley: Path to Effective Long CoT Training for Small Language Models

Paper • 2506.07712 • Published about 1 month ago • 18

MiniCPM4: Ultra-Efficient LLMs on End Devices

Paper • 2506.07900 • Published about 1 month ago • 82

upvoted a paper about 1 month ago

The Common Pile v0.1: An 8TB Dataset of Public Domain and Openly Licensed Text

Paper • 2506.05209 • Published Jun 5 • 42