0

Medconceptsqa RL Env (Medarc)

Fresh

Your environment description here

Type
RL Env
Publisher
Medarc
Runtime
single-turn
License
unknown
Size
v0.1.1
Published
Dec 2025

Cite

Notes

Only stored in your browser.

medconceptsqa

Overview

  • Environment ID: medconceptsqa
  • Short description: MedConcepts QA - an MCQ dataset involving medical codes.
  • Tags: medical, clinical, single-turn, multiple-choice, classification, test

Datasets

  • Primary dataset(s): medconceptsqa
  • Source links: Paper, Github, HF Dataset
  • Split sizes: 60 (dev / few-shot), 820k (test)

Task

  • Type: single-turn
  • Rubric overview: Binary scoring based on correct answer choice

Quickstart

Run an evaluation with default settings:

prime eval run medconceptsqa -m "openai/gpt-5-mini" -n 5 -s

Configure model and sampling:

medarc-eval medconceptsqa -m "openai/gpt-5-mini" -n 20 --num-few-shot 4

Notes:

  • Use direct environment flags with medarc-eval (for example, --split validation or --judge-model gpt-5-mini).

Environment Arguments

ArgTypeDefaultDescription
num_few_shotint0Number of few-shot examples to include in the prompt
use_thinkboolFalseWhether to use <think>...</think> formatting with ThinkParser

Metrics

MetricMeaning
accuracyExact match on target answer

Authors

This environment has been put together by:

Anish Mahishi - (@macandro96)