Sorry if I’m misunderstanding something here, but ...
# proj-codescout
r
Sorry if I’m misunderstanding something here, but which dataset was used to generate the table in https://github.com/OpenHands/agentic-code-search-oss/issues/24? I followed the workflow described in README_verifiers.md to run the evals, but it seems that the swebench_lite dataset does not directly provide the fields needed to evaluate Module Loc (F1) and Function/Entity Loc (F1). Am I missing something, or is there an additional preprocessing step or dataset involved?