diff --git a/keybert/_mmr.py b/keybert/_mmr.py
index ae6b22eb..c11db0f9 100644
--- a/keybert/_mmr.py
+++ b/keybert/_mmr.py
@@ -41,6 +41,7 @@ def mmr(
     # Initialize candidates and already choose best keyword/keyphras
     keywords_idx = [np.argmax(word_doc_similarity)]
     candidates_idx = [i for i in range(len(words)) if i != keywords_idx[0]]
+    words_mmr = [(1 - diversity) * np.max(word_doc_similarity)]
 
     for _ in range(min(top_n - 1, len(words) - 1)):
         # Extract similarities within candidates and
@@ -55,15 +56,14 @@ def mmr(
             1 - diversity
         ) * candidate_similarities - diversity * target_similarities.reshape(-1, 1)
         mmr_idx = candidates_idx[np.argmax(mmr)]
+        words_mmr.append(np.max(mmr))
 
         # Update keywords & candidates
         keywords_idx.append(mmr_idx)
         candidates_idx.remove(mmr_idx)
 
     # Extract and sort keywords in descending similarity
-    keywords = [
-        (words[idx], round(float(word_doc_similarity.reshape(1, -1)[0][idx]), 4))
-        for idx in keywords_idx
-    ]
+    keywords = [(words[keywords_idx[i]], round(float(words_mmr[i]), 4))
+                for i in range(len(keywords_idx))]
     keywords = sorted(keywords, key=itemgetter(1), reverse=True)
     return keywords