LLM Output Table appears truncated

I’m attempting to Extract Entities from a variety of pdf’s that are multiple pages. The input table has 78 rows, however, the output table consistently has 20 rows.

I’ve attempted to include errors and there are none returned. Changing the max tokens also seems to have no effect.

Hey! Just a few follow up questions:

  1. So the input has 78 rows, after you save and deploy it only returns 20 rows but you’re expecting the full 78
  2. Are you doing anything else in this pipeline? Any joins/filters or is it just a use llm node

It is just an LLM node with a single input that is 78 rows.

Are you able to share the prompt? Also, confirming this is after you save the pipeline and deploy it (not on preview?)