REBEL: Reinforcement Learning via Regressing Relative Rewards
Paper • 2404.16767 • Published • 2
REBEL: Reinforcement Learning via Regressing Relative Reward
-
-
Cornell-AGI/REBEL-Llama-3-Armo-iter_1
8B • Updated • 2 • 1 -
Cornell-AGI/REBEL-Llama-3-Armo-iter_2
8B • Updated • 2 • 1 -
Cornell-AGI/REBEL-Llama-3-Armo-iter_3
8B • Updated • 3 • 2 -
Cornell-AGI/Ultrafeedback-Llama-3-Armo-iter_1
Viewer • Updated • 56.1k • 7 -
Cornell-AGI/Ultrafeedback-Llama-3-Armo-iter_2
Viewer • Updated • 55.1k • 7 -
Cornell-AGI/Ultrafeedback-Llama-3-Armo-iter_3
Viewer • Updated • 44.6k • 7 • 1 -
Cornell-AGI/REBEL-Llama-3
Text Generation • Updated • 6 • 1 -
Cornell-AGI/REBEL-Llama-3-epoch_2
Text Generation • Updated • 12 • 3 -
Cornell-AGI/REBEL-OpenChat-3.5
Text Generation • Updated • 14 • 1