Ken Tsui

kenhktsui

AI & ML interests

ML Engineer Lead. Researcher on Small Language Model - Building Classifiers to Find High Quality Data/ Reasoning Benchmark/ Synthetic Data

Recent Activity

liked a model about 22 hours ago

google/gemma-3-27b-it

liked a model 2 days ago

MCG-NJU/videomae-base

liked a dataset 5 days ago

ontocord/lighteval_MATH-Hard

View all activity

Organizations

kenhktsui's activity

upvoted an article about 1 month ago

Article

Open-source DeepResearch – Freeing our search agents

Feb 4

• 1.16k

upvoted a paper about 1 month ago

ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoning

Paper • 2502.01100 • Published Feb 3 • 17

upvoted a paper 2 months ago

rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking

Paper • 2501.04519 • Published Jan 8 • 263

upvoted 2 papers 3 months ago

Byte Latent Transformer: Patches Scale Better Than Tokens

Paper • 2412.09871 • Published Dec 13, 2024 • 93

ProcessBench: Identifying Process Errors in Mathematical Reasoning

Paper • 2412.06559 • Published Dec 9, 2024 • 80

upvoted 4 articles 7 months ago

Article

How NuminaMath Won the 1st AIMO Progress Prize

Jul 11, 2024

• 118

Article

Merge Large Language Models with mergekit

•

Jan 9, 2024

• 102

Article

Llama-3.1-Storm-8B: Improved SLM with Self-Curation + Model Merging

•

Aug 19, 2024

• 77

Article

∞🧙🏼‍♂️AnyClassifier - Generating Synthetic Data For Text Classification

•

Aug 19, 2024

• 8

upvoted an article 8 months ago

Article

ZebraLogic: Benchmarking the Logical Reasoning Ability of Language Models

•

Jul 27, 2024

• 31

upvoted a paper 8 months ago

Scaling Laws with Vocabulary: Larger Models Deserve Larger Vocabularies

Paper • 2407.13623 • Published Jul 18, 2024 • 55

upvoted an article 9 months ago

Article

Low Latency CPU Based Educational Value Classifier With Generic Educational Value

•

Jun 12, 2024

• 9

upvoted 2 papers 10 months ago

Direct Nash Optimization: Teaching Language Models to Self-Improve with General Preferences

Paper • 2404.03715 • Published Apr 4, 2024 • 61

LoRA Learns Less and Forgets Less

Paper • 2405.09673 • Published May 15, 2024 • 88

upvoted a paper 11 months ago

InstantFamily: Masked Attention for Zero-shot Multi-ID Image Generation

Paper • 2404.19427 • Published Apr 30, 2024 • 72

upvoted a paper 12 months ago

Aurora-M: The First Open Source Multilingual Language Model Red-teamed according to the U.S. Executive Order

Paper • 2404.00399 • Published Mar 30, 2024 • 42

upvoted a collection about 1 year ago

Tiny Series

Collection

Tiny datasets that empower the foundation of Small Language Model! • 11 items • Updated Jan 26, 2024 • 36