Baidu's Improving Retrieval Augmented Language Model with Self-Reasoning

66 points by a-s-k-af 1 year ago | 4 comments

Can anyone explain what is gained by training a model? Why not use the foundational LLM for the relevance, evidence, and trajectory processes?

a-s-k-af 1 year ago | |

I assume you are referring to fine tuning a model here?

Tostino 1 year ago | | |

You could also just continue pre-training of an existing foundation model. Would still be cheaper by not starting from zero.