If you run one of the distill versions in something like LM Studio it’s very easy to edit. But the replies from those models isn’t half as good as the full R1, but still remarkably better then anything I’ve run locally before.