After finishing the data creation or selection part (see step A to F in text simplification workflow; see Figure 2.1 in Subsection 2.2.2), a text simplification model can be trained (see step G) and evaluated on this data (see step H). After the first evaluation in the development phase against automatic metrics, the parameters of the models will be tuned to achieve better performance in the given task, here text simplification. To better understand TS models’ performance, I am explaining in this section how and why TS models are evaluated.