Finalize markdown-table RAG LLM strategy

jay-joshy / question_cli

0 stars 0 forks source link

Finalize markdown-table RAG LLM strategy #7

Closed jay-joshy closed 2 months ago

jay-joshy commented 3 months ago

When prompting the LLM agent A to review the retrieved context, automatically discard parts of the generated md table that do not have a high rating. Should improve quality of context fed to agent B

jay-joshy commented 2 months ago

pivoted to pydantic for more consistent outputs