Chunyuan Deng's picture

Chunyuan Deng

CharlesDDDD

·

https://charlesdddd.github.io/

AI & ML interests

Architecheture, Interpretability.

Recent Activity

updated a model 2 days ago

CharlesDDDD/looped_600M_gdn_4to1

published a model 2 days ago

CharlesDDDD/looped_600M_gdn_4to1

updated a model 2 days ago

CharlesDDDD/looped_window_attnetion_1B_formal

View all activity

Organizations

Collections 2

Papers 4

arxiv:2603.03583

arxiv:2502.13131

arxiv:2410.15580

arxiv:2311.09783

models 8

CharlesDDDD/looped_window_attnetion_1B_formal

Updated 2 days ago • 4

CharlesDDDD/looped_600M_gdn_4to1

Updated 2 days ago • 5

CharlesDDDD/looped_transformer_loop_count_4

Updated Feb 5 • 1

CharlesDDDD/looped_600M_4to1_looped1_window_128

CharlesDDDD/looped_600M_bookend_2_looped4_window_128

CharlesDDDD/llama_1.3B_100bt_formal

Updated Feb 4 • 2

CharlesDDDD/llama_0.6B_100bt_formal

Updated Feb 4 • 2

CharlesDDDD/looped_window_600M_full_512_256_128_weighted_loss

datasets 0

None public yet