70m Pythia model after SFT on the AlpacaFarm dataset 'sft' split.
Model used as a base for reward models in 'Reward Model Ensembles Mitigate Overoptimization'
- Downloads last month
- 1,379
70m Pythia model after SFT on the AlpacaFarm dataset 'sft' split.
Model used as a base for reward models in 'Reward Model Ensembles Mitigate Overoptimization'