MediQ

[warner-benjamin]

Implement verifiers env for MediQ

Implement a verifiers environment for the MediQ medical dataset. Use the authors’ question, system, and LLM-as-judge prompts if provided. Only create a training dataset if a training split exists.

Paper: https://arxiv.org/abs/2406.00922
Dataset / Project: https://github.com/stellalisy/mediQ

Format & Env

• Multi-turn interactive reasoning: vf.MultiTurnEnv
• Open-ended responses: vf.JudgeRubric()
• Supports follow-up questioning before final answers

[mkieffer1107]

I can work on this one