Lukas

sirluk

·

AI & ML interests

None yet

Recent Activity

upvoted an article 4 days ago

Deriving the PPO Loss from First Principles

published an article 14 days ago

Sliding window beats linear attention, but only on short answers

commentedon their article 16 days ago

Efficient LLM Pretraining: Packed Sequences and Masked Attention

View all activity

Organizations

Articles 3

Article

6

Sliding window beats linear attention, but only on short answers

Article

74

Efficient LLM Pretraining: Packed Sequences and Masked Attention

View all Articles

Papers 3

arxiv:2607.05061

arxiv:2603.15590

arxiv:2410.07170

spaces 1

Trl Lora Without Regret Smollm3

Display tracking information

models 1

sirluk/Qwen3-8B-KVpop-4x

Text Generation • 8B • Updated Jul 13 • 18 • 3

datasets 0

None public yet