A dataset of sentences and the questions that they answer.
Propositions were randomly sampled from agentlans/wikipedia-propositions
Then rewritten as questions using google/gemma-3-12b-it
Some question-answer pairs sound unnatural
Lacks context when processing single sentences
