Open
Description
Description
In order to evaluate the efficacy of the system, there will need to be a pool of curated questions and answers that can be used to test the system.
Acceptance Criteria
- 200 question/answer pairs for the AI legislation that the system will work with
- Some questions should be specific to individual legislation
- Some questions should involve comparing multiple legislation
- Some questions should involve asking about specific state(s)
Developer notes
LLMs should be leveraged in order to rapidly generate the question answer pairs
- Summarizing large blocks of texts into an answer
- Scanning documents to generate questions
Metadata
Metadata
Assignees
Labels
No labels