diff --git a/README.md b/README.md
index d0e2543..c52c780 100644
--- a/README.md
+++ b/README.md
@@ -59,12 +59,12 @@ Comparing with cloning or copying into `smallmap::Map`.
 
 Largely there are the same performance penalties as the above table, with very minor differences.
 
-| Benchmark          | Tests                                         | Result          |
-|--------------------|-----------------------------------------------|-----------------|
-| owning_strings     | Inserts `String` into `HashSet` by cloning    | ~3,096 ns/iter  |
-| non_owning_strings | Inserts `str` into `HashRefSet` by reference  | ~47,302 ns/iter |
-| owning_ints        | Inserts `u32` into `HashSet` by copy          | ~316 ns/iter    |
-| non_owning_ints    | Inserts `&u32` into `HashRefSet` by reference | ~30,046 ns/iter |
+| Benchmark          | Tests                                          | Result          |
+|--------------------|------------------------------------------------|-----------------|
+| owning_strings     | Inserts `String` into `SmallMap` by cloning    | ~3,096 ns/iter  |
+| non_owning_strings | Inserts `str` into `SmallRefMap` by reference  | ~47,302 ns/iter |
+| owning_ints        | Inserts `u32` into `SmallMap` by copy          | ~316 ns/iter    |
+| non_owning_ints    | Inserts `&u32` into `SmallRefMap` by reference | ~30,046 ns/iter |
 
 Each page of the `SmallRefMap` will consume at least 16kb of memory however.
 This may not be very desireable, but is still an available feature.