For fine-tuning SFT, we focused on extending machine translation capabilities by prepping data focused on post-editing, error detection, and terminology.
Original titleFor fine-tuning SFT, we focused on extending machine translation capabilities by prepping data focused on post-editing, error detection, ...
AISummary
We also went through multiple steps of reinforcement learning and DPO targeted towards errors the model exhibited. (2/6)
Source: Cohere · x.com