Replies: 2 comments
-
Do you have a specific example? You can ask the LLM to not include any quotes in the output JSON to save around 10% of tokens as well |
Beta Was this translation helpful? Give feedback.
-
Thank you for the quick answer. Any use case that involves tabular data with large number of rows comes under this requirement For example, I have a csv list like below (Facebook insights fields). For each of the field , i need the llm to assign a user friendly name, and possible data type . (The below list may be small and may not be a true representation of the use case. But there could be similar cases where there are thousands of rows, in which case json becomes not as useful as csv)
|
Beta Was this translation helpful? Give feedback.
-
Thank you for this library. Wanted to ask a quick question.
JSON format is token intensive. Many a times, letting the llm output data in csv format is needed particularly when dealing with tabular data. Does BAML support that ? Or is it outside the scope of this project
Beta Was this translation helpful? Give feedback.
All reactions