Teaching Pretrained Language Models to Think Deeper with Retrofitted Recurrence
Paper • 2511.07384 • Published • 21
Retrofitting Recurrence
-
-
smcleish/Recurrent-Llama-3.2-train-recurrence-32
Text Generation • 1B • Updated • 212 • 1 -
smcleish/Recurrent-Llama-3.2-train-recurrence-16
Text Generation • 1B • Updated • 3.18k -
smcleish/Recurrent-Llama-3.2-train-recurrence-8
Text Generation • 1B • Updated • 60 -
smcleish/Recurrent-Llama-3.2-train-recurrence-4
Text Generation • 1B • Updated • 69 -
smcleish/Recurrent-TinyLlama-3T-train-recurrence-32
Text Generation • 0.8B • Updated • 67 • 2 -
smcleish/Recurrent-TinyLlama-3T-train-recurrence-16
Text Generation • 0.8B • Updated • 9 • 1 -
smcleish/Recurrent-TinyLlama-3T-train-recurrence-8
Text Generation • 0.8B • Updated • 65 -
smcleish/Recurrent-TinyLlama-3T-train-recurrence-4
Text Generation • 0.8B • Updated • 43 -
smcleish/Recurrent-OLMo-2-0425-train-recurrence-32
Text Generation • 1B • Updated • 205 • 3 -
smcleish/Recurrent-OLMo-2-0425-train-recurrence-16
Text Generation • 1B • Updated • 8.95k -
smcleish/Recurrent-OLMo-2-0425-train-recurrence-8
Text Generation • 1B • Updated • 33 -
smcleish/Recurrent-OLMo-2-0425-train-recurrence-4
Text Generation • 1B • Updated • 14 • 1 -
smcleish/Recurrent-TinyLlama-3T-train-recurrence-4-single-phase
Text Generation • 0.8B • Updated • 15 -
smcleish/Recurrent-TinyLlama-3T-train-recurrence-4-two-phase
Text Generation • 0.8B • Updated • 10 -
smcleish/Recurrent-Llama-3.2-untrained
Text Generation • 1B • Updated • 9 -
smcleish/Recurrent-TinyLlama-3T-untrained
Text Generation • 0.8B • Updated • 65 -
smcleish/Recurrent-OLMo-2-0425-untrained
Text Generation • 1B • Updated • 12 -
smcleish/Recurrent-Llama-3.2-2-4-2-untrained
Text Generation • 1B • Updated • 14 • 1 -
smcleish/retrofitting-llama-fineweb-edu-tokenized
Viewer • Updated • 332M • 59 • 1 -
smcleish/Llama-3.2-non-recurrent-posttrained
Text Generation • 1B • Updated • 60