Peter Szemraj PRO

pszemraj's activity

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

upvoted a paper about 14 hours ago

OLMoTrace: Tracing Language Model Outputs Back to Trillions of Training Tokens

Paper • 2504.07096 • Published 5 days ago • 64

upvoted a collection 5 days ago

SuperBPE

Collection

SuperBPE tokenizers and models trained with them • 8 items • Updated 5 days ago • 14

upvoted 2 papers 11 days ago

PaperBench: Evaluating AI's Ability to Replicate AI Research

Paper • 2504.01848 • Published 12 days ago • 34

Efficient Inference for Large Reasoning Models: A Survey

Paper • 2503.23077 • Published 16 days ago • 45

upvoted 2 papers 12 days ago

A Survey of Efficient Reasoning for Large Reasoning Models: Language, Multimodality, and Beyond

Paper • 2503.21614 • Published 18 days ago • 39

Multi-Token Attention

Paper • 2504.00927 • Published 13 days ago • 42

upvoted a paper 20 days ago

I Have Covered All the Bases Here: Interpreting Reasoning Features in Large Language Models via Sparse Autoencoders

Paper • 2503.18878 • Published 21 days ago • 113

upvoted 3 papers 23 days ago

Survey on Evaluation of LLM-based Agents

Paper • 2503.16416 • Published 25 days ago • 84

One-Step Residual Shifting Diffusion for Image Super-Resolution via Distillation

Paper • 2503.13358 • Published 28 days ago • 95

Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models

Paper • 2503.16419 • Published 25 days ago • 68

upvoted 2 papers 25 days ago

Optimizing Decomposition for Optimal Claim Verification

Paper • 2503.15354 • Published 26 days ago • 18

φ-Decoding: Adaptive Foresight Sampling for Balanced Inference-Time Exploration and Exploitation

Paper • 2503.13288 • Published 28 days ago • 49

upvoted a paper 28 days ago

Transformers without Normalization

Paper • 2503.10622 • Published Mar 13 • 155

upvoted 6 papers about 1 month ago

Implicit Reasoning in Transformers is Reasoning through Shortcuts

Paper • 2503.07604 • Published Mar 10 • 21

WritingBench: A Comprehensive Benchmark for Generative Writing

Paper • 2503.05244 • Published Mar 7 • 17

DistiLLM-2: A Contrastive Approach Boosts the Distillation of LLMs

Paper • 2503.07067 • Published Mar 10 • 31

SEAP: Training-free Sparse Expert Activation Pruning Unlock the Brainpower of Large Language Models

Paper • 2503.07605 • Published Mar 10 • 66

Feature-Level Insights into Artificial Text Detection with Sparse Autoencoders

Paper • 2503.03601 • Published Mar 5 • 227

HybridNorm: Towards Stable and Efficient Transformer Training via Hybrid Normalization

Paper • 2503.04598 • Published Mar 6 • 18