Yixuan Wang

LuckyOrz

6 4

·

LuckySJTU

AI & ML interests

LLM pretraining, hierarchical model, efficient LLM

Recent Activity

upvoted a paper 3 days ago

Memory Decoder at Scale: A Pretrained, Parametric Long-Term Memory

upvoted a paper 14 days ago

xHC: Expanded Hyper-Connections

upvoted a paper about 1 month ago

Information-Aware KV Cache Compression for Long Reasoning

View all activity

Organizations

upvoted a paper 3 days ago

Memory Decoder at Scale: A Pretrained, Parametric Long-Term Memory

Paper • 2607.27919 • Published 4 days ago • 49

upvoted a paper 14 days ago

xHC: Expanded Hyper-Connections

Paper • 2607.14530 • Published 18 days ago • 56

upvoted a paper about 1 month ago

Information-Aware KV Cache Compression for Long Reasoning

Paper • 2606.26875 • Published Jun 25 • 12

upvoted a collection 6 months ago

ConceptLM

The Collection of ConceptLM models and data, ConceptLM integrates Next-Concept-Prediction and Next-Token-Prediction. • 2 items • Updated Feb 3 • 5

upvoted a paper 11 months ago

FlowRL: Matching Reward Distributions for LLM Reasoning

Paper • 2509.15207 • Published Sep 18, 2025 • 119

upvoted a paper over 1 year ago

How to Synthesize Text Data without Model Collapse?

Paper • 2412.14689 • Published Dec 19, 2024 • 52