diff --git a/rfcs/20230622-quantized-reduction.md b/rfcs/20230622-quantized-reduction.md
index e0418cddb52..e06bfe24972 100644
--- a/rfcs/20230622-quantized-reduction.md
+++ b/rfcs/20230622-quantized-reduction.md
@@ -41,7 +41,7 @@ The RFC introduces the following proposal, emerged out of discussion in the
 [thread](https://github.com/openxla/stablehlo/pull/1538#issuecomment-1599476906)
 , along with their tradeoffs.
 
-The proposal allows the reducer block to express the computation in a different
+The proposal optionally allows the reducer block to express the computation in a different
 element type (preferably wider accumulation type) than the one used in reduce
 op's ops arguments and return type. For illustrative purposes, in the following
 example, the operand element type
@@ -71,11 +71,14 @@ block return (`tensor<!quant.uniform<i32:f32, accum_scale:accum_zp>>`).
 
 ### Semantics
 
-If (1) the input operand type is different from the reduction block
-argument type or (2) the op result type is different from the reduction block
-return type, there will be implicit type conversion defined by either
+If (1) the input operand type is different from the reduction block argument
+type or (2) the op result type is different from the reduction block return
+type, there will be implicit type conversion defined by either
 `stablehlo.convert`, `stablehlo.uniform_quantize`, or
-`stablehlo.uniform_dequantize`. For example,
+`stablehlo.uniform_dequantize`. When the types are not differnet, i.e., when (1)
+and (2) does not hold true, then no implicit convertion is needed.
+
+For example,
 
  | Implicit type conversion op       | element type of operand or block return | element type of block argument or op return |
  |-----------------------------------|-----------------------------------------|---------------------------------------------|