How can I simplify or clean up this anagram method?

Question

I have a method here that takes an array of strings and groups the ones that are anagrams of each other together, with each group forming a sub-array of the main anagr

Answer 1

test_words.group_by{|w| w.each_char.sort}.values

would give

[
  ["cars", "racs", "scar"],
  ["for"],
  ["potatoes"],
  ["four"],
  ["creams", "scream"]
]

Answer 2

I modified sawa's answer slightly in order to ignore case and make sure there's no duplicate values:

test_words.group_by{|w| w.downcase.each_char.sort}.values.each{|v| v.uniq!}

I realize this will still give duplicates in the output if the words have characters with different cases, but that's fine for my purposes. Now I'm all sorted, thanks!