Task dataset for a verifiers RL / eval environment, in the shape used by the
Prime Intellect Environments Hub.
22 tasks about the verifiers library itself — the library every environment in that program
is written against. The skill tested is the one a contributor actually needs: parsing a model's
output, writing reward functions, and composing them into a Rubric whose weighting does what
you meant.
parsers
7
Parser, XMLParser… See the full description on the dataset page:
https://huggingface.co/datasets/eltociear/verifiers-tasks-v1.
