Skip to content

Dataset eval helper #44

@deadbits

Description

@deadbits

Vigil should have a helper utility to assist users in evaluating datasets against the detection mechanisms and return true positives, false positives, false negatives, accuracy score, etc.

I'm thinking a simple CSV of prompts, classifications, and categories.

Metadata

Metadata

Assignees

Labels

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions