A collection of the released datasets, tokenizer training split and evaluated benchmarks for the ObscuraCoder paper.