You can use apply, like this.
In [13]: df[df['col'].apply(lambda x: 'b' in x)]
Out[13]:
col
0 [a, b]
2 [a, b, c]
Although generally, storing lists in a DataFrame
is a bit awkward – you might find some different representation (columns for each element in the list, MultiIndex, etc) that is easier to work with.