MLI will by default perform sampling if your dataset has more than 100K rows. You can disable this or set the sampling limit to a higher number through expert settings or config.toml. To increase the sampling limit:
# When number of rows are above this limit sample for MLI for scoring UI data mli_sample_above_for_scoring = 2000000 # When number of rows are above this limit sample for MLI for training surrogate models mli_sample_above_for_training = 200000 # When sample for MLI how many rows to sample mli_sample_size = 200000
This will set the limit to 2000000. To disable MLI sampling:
# not only sample training, but also sample scoring mli_sample_training = false
You will have to re-run your experiment after changing these configurations.