Weak-to-Strong Extrapolation Expedites Alignment
Paper
•
2404.16792
•
Published
•
11
Model Extrapolation Expedites Alignment
Better aligned models obtained by model extrapolation (ExPO)
-
-
chujiezheng/Mistral7B-PairRM-SPPO-ExPO
Text Generation • Updated • 5.92k -
chujiezheng/LLaMA3-iterative-DPO-final-ExPO
Text Generation • Updated • 5.19k • 2 -
chujiezheng/Llama3-8B-Chinese-Chat-ExPO
Text Generation • Updated • 5.21k • 1 -
chujiezheng/tulu-2-dpo-70b-ExPO
Text Generation • Updated • 5.19k -
chujiezheng/Smaug-Llama-3-70B-Instruct-ExPO
Text Generation • Updated • 5.2k • • 2 -
chujiezheng/Starling-LM-7B-beta-ExPO
Text Generation • Updated • 4.98k • 2 -
chujiezheng/Starling-LM-7B-alpha-ExPO
Text Generation • Updated • 4.95k -
chujiezheng/Llama3-70B-Chinese-Chat-ExPO
Text Generation • Updated • 5.15k -
chujiezheng/internlm2-chat-20b-ExPO
Text Generation • Updated • 10k • 1 -
chujiezheng/internlm2-chat-7b-ExPO
Text Generation • Updated • 10k -
chujiezheng/Smaug-34B-v0.1-ExPO
Text Generation • Updated • 5.16k -
chujiezheng/tulu-2-dpo-13b-ExPO
Text Generation • Updated • 14 -
chujiezheng/Llama-3-Instruct-8B-SimPO-ExPO
Text Generation • Updated • 22 • 16 -
chujiezheng/tulu-2-dpo-7b-ExPO
Text Generation • Updated • 17 -
chujiezheng/Snorkel-Mistral-PairRM-DPO-ExPO
Text Generation • Updated • 11 -
chujiezheng/neo_7b_instruct_v0.1-ExPO
Text Generation • Updated • 11 -
chujiezheng/zephyr-7b-beta-ExPO
Text Generation • Updated • 8 -
chujiezheng/zephyr-7b-alpha-ExPO
Text Generation • Updated • 12 • 1 -
chujiezheng/zephyr-7b-dpo-full-ExPO
Text Generation • Updated • 16 -
chujiezheng/internlm2-chat-1_8b-ExPO
Text Generation • Updated • 19 • 1