Baidu's Improving Retrieval Augmented Language Model with Self-Reasoning (opens in new tab)

(arxiv.org)

66 pointsa-s-k-af1y ago4 comments

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?

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

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

j / k navigate · click thread line to collapse

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

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

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

j / k navigate · click thread line to collapse