File-level Truncation

[nolanlwin]

Feature Description

Currently, when L2M sends a file to LLM, it sends the entire file (even if a user request is specific to ceterin LOCs). It should be able to truncate the particular LOCs before sending it so that it will be more token and time efficient.

Implementation Ideas

TBD

Deliverables

TBD