Mopsa-format records with a natural-language rewrite generated from the concrete execution trace. The projected trace_seps field stays in the same shape as the existing natural dataset, while the raw execution-trace fields are preserved alongside it.
Columns:
code: original source code.
trace_seps: projected execution-trace segments in the usual generic format.
dataset: source dataset name.
id: example identifier.
output_varname: final printed variable name.… See the full description on the dataset page:
https://huggingface.co/datasets/giovannidemuri/mopsa-dataset-natural-execution.
