A researcher observes a curved (non-linear) relationship between two variables. She transforms the predictor using $x' = \ln(x)$ and fits the model $\hat{y} = 2.0 \cdot \ln(x) + 5$ with $R^2 = 0.92$. What does this transformation suggest about the original relationship?