Assuming you mean n=100, then Well for this I would choose large absolute values of your DV and replace the data values with them. For a larger study I would google and find lots of programs that do this in your fav stat packages. See the link: generating outliers in r - Bing

remember to sample from something like a Cauchy or Laplace distribution. Best wishes, David Booth

You haven't defined what you mean by outlier (e.g., univariate vs multivariate, what the assumed underlining processes are for outlying and non-outlying data, etc). And as David Eugene Booth says, it is not clear if you mean you have a set of variables, but let's assume that it is a data set with 100 rows, and call this X. Let whatever you define outlier as by a function you write called out. Then simply apply out to those rows.

Given you haven't supplied adequate information, I'd recommend deleting your question and starting over.

See these links:

https://www.researchgate.net/post/What_are_the_best_methods_to_generate_artificial_outliers_in_one_class_classification_task

https://towardsdatascience.com/a-brief-overview-of-outlier-detection-techniques-1e0b2c19e561

https://www.thoughtco.com/what-is-an-outlier-3126227

https://towardsdatascience.com/outliers-analysis-a-quick-guide-to-the-different-types-of-outliers-e41de37e6bf6