Method clean_cpp

lm-eval-harness/lm_eval/decontamination/janitor.py:200–204

(self, dirty_string: str)

        )

    def clean_cpp(self, dirty_string: str) -> List[str]:
        contamination_indices = janitor_util.clean_ngram_with_indices(
            dirty_string, self.delete_chars, self.ngram_n
        )
        return self._split_chunks(dirty_string, contamination_indices)

    ##############
    # Slow python