in the article they talk about syntactically invalid smile formula (mismatch parenthesis and cie, check figure 1b), impossible molecules can be expressed with s

in the article they talk about syntactically invalid smile formula (mismatch parenthesis and cie, check figure 1b), impossible molecules can be expressed with syntactically valid SMILE. But nonetheless your explanation appears valid to my non chemist eyes. from the paper: Structural biases limit generalization An ideal generative model would sample evenly from the chemical space surrounding the molecules in the training set. The observation of structural biases in the outputs of language models trained on SELFIES is at odds with this goal.