Comment by gbasin

For large models that are well-trained with large context support, ICL seems to work about as well as FT