The second-language English writing domain is built from EFCAMDAT, which is distributed under an academic user agreement. That agreement licenses the corpus to the user rather than making it redistributable, so this directory starts empty and you fill it from your own copy.

The cleaned error-coded subcorpus of EFCAMDAT (Öksüz et al., 2025). This is the release that carries span-level teacher corrections, which the multi-turn records need. The other EFCAMDAT releases do not carry them and will not work.

Request access from the maintainers of the corpus. The build reads one file:

data/l2/raw/ef_POStagged_original_corrected.csv

python -m studentsim.data.l2.build