diff --git a/mlir/include/mlir/Dialect/Vector/Utils/VectorUtils.h b/mlir/include/mlir/Dialect/Vector/Utils/VectorUtils.h
index 8a57c6094c41c..30e21c64c2731 100644
--- a/mlir/include/mlir/Dialect/Vector/Utils/VectorUtils.h
+++ b/mlir/include/mlir/Dialect/Vector/Utils/VectorUtils.h
@@ -185,8 +185,8 @@ bool isLinearizableVector(VectorType type);
 /// mask is created on the read, if use of mask is specified or the bounds on a
 /// dimension are different.
 ///
-/// `useInBoundsInsteadOfMasking` if false, the inBoundsVal values are set
-/// properly, based on
+/// `useInBoundsInsteadOfMasking` if true, the inBoundsVal values are set
+///   properly, based on
 ///   the rank dimensions of the source and destination tensors. And that is
 ///   what determines if masking is done.
 ///
@@ -194,7 +194,7 @@ bool isLinearizableVector(VectorType type);
 /// for each dimension of the passed in tensor.
 Value createReadOrMaskedRead(OpBuilder &builder, Location loc, Value source,
                              ArrayRef<int64_t> readShape, Value padValue,
-                             bool useInBoundsInsteadOfMasking = true);
+                             bool useInBoundsInsteadOfMasking = false);
 
 /// Returns success if `inputVectorSizes` is a valid masking configuraion for
 /// given `shape`, i.e., it meets:
diff --git a/mlir/lib/Dialect/Linalg/Transforms/Vectorization.cpp b/mlir/lib/Dialect/Linalg/Transforms/Vectorization.cpp
index e836f0dc63b4f..be1b602783fe9 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/Vectorization.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/Vectorization.cpp
@@ -1499,11 +1499,11 @@ vectorizeAsTensorPackOp(RewriterBase &rewriter, tensor::PackOp packOp,
   // If the input vector sizes are not provided, then the vector sizes are
   // determined by the result tensor shape. In case the vector sizes aren't
   // provided, we update the inBounds attribute instead of masking.
-  bool useInBoundsInsteadOfMasking = true;
+  bool useInBoundsInsteadOfMasking = false;
   if (inputVectorSizes.empty()) {
     ArrayRef<int64_t> resultTensorShape = packOp.getDestType().getShape();
     inputVectorSizes = resultTensorShape.take_front(packOp.getSourceRank());
-    useInBoundsInsteadOfMasking = false;
+    useInBoundsInsteadOfMasking = true;
   }
 
   // Create masked TransferReadOp.
diff --git a/mlir/lib/Dialect/Vector/Utils/VectorUtils.cpp b/mlir/lib/Dialect/Vector/Utils/VectorUtils.cpp
index fcaf1ec944b47..f9ddaa08a1910 100644
--- a/mlir/lib/Dialect/Vector/Utils/VectorUtils.cpp
+++ b/mlir/lib/Dialect/Vector/Utils/VectorUtils.cpp
@@ -345,7 +345,7 @@ Value vector::createReadOrMaskedRead(OpBuilder &builder, Location loc,
   int64_t readRank = readShape.size();
   auto zero = builder.create<arith::ConstantIndexOp>(loc, 0);
   SmallVector<bool> inBoundsVal(readRank, true);
-  if (!useInBoundsInsteadOfMasking) {
+  if (useInBoundsInsteadOfMasking) {
     // Update the inBounds attribute.
     for (unsigned i = 0; i < readRank; i++)
       inBoundsVal[i] = (sourceShape[i] == readShape[i]) &&