Predicted probabilities that a molecule elicits a response are evaluated using the log loss metric.
Log loss is defined as:
$$\text{log loss}=-\frac{1}{N}\sum_{i=1}^Ny_i\log\left(\hat{y_i}\right)+\left(1-y_i\right)\log\left(1-\hat{y_i}\right),$$
where
\\(N\\)
is the number of samples,
\\(\text{log}\\)
is the natural logarithm,
\\(\hat{y_i}\\)
is the posterior probability that the
\\(i^{th}\\)
sample elicited a response, and
\\(y_i\\)
is the ground truth
(\\(y_i=1\\)
means the molecule elicited a response,
\\(y_i=0\\)
means that it did not).