diff --git a/src/metric/auc.cc b/src/metric/auc.cc
index 9184223da..f3613228f 100644
--- a/src/metric/auc.cc
+++ b/src/metric/auc.cc
@@ -126,8 +126,8 @@ float MultiClassOVR(std::vector<float> const& predts, MetaInfo const& info) {
   for (size_t c = 0; c < n_classes; ++c) {
     if (local_area[c] != 0) {
       // normalize and weight it by prevalence.  After allreduce, `local_area` means the
-      // total covered area (not area under curve, rather it's the accessible are for each
-      // worker) for each class.
+      // total covered area (not area under curve, rather it's the accessible area for
+      // each worker) for each class.
       auc_sum += auc[c] / local_area[c] * tp[c];
       tp_sum += tp[c];
     } else {
@@ -310,7 +310,7 @@ class EvalAUC : public Metric {
   }
 };
 
-XGBOOST_REGISTER_METRIC(EvalBinaryAUC, "auc")
+XGBOOST_REGISTER_METRIC(EvalAUC, "auc")
 .describe("Receiver Operating Characteristic Area Under the Curve.")
 .set_body([](const char*) { return new EvalAUC(); });
 
diff --git a/src/metric/auc.cu b/src/metric/auc.cu
index 433d84710..615dad5ae 100644
--- a/src/metric/auc.cu
+++ b/src/metric/auc.cu
@@ -74,7 +74,7 @@ struct DeviceAUCCache {
  *
  * - Run scan to obtain TP/FP values, which are right coordinates of trapesoid.
  * - Find distinct prediction values and get the corresponding FP_PREV/TP_PREV value,
- *   which are left coordinates of trapesoid.
+ *   which are left coordinates of trapesoids.
  * - Reduce the scan array into 1 AUC value.
  */
 std::tuple<float, float, float>
@@ -151,7 +151,7 @@ GPUBinaryAUC(common::Span<float const> predts, MetaInfo const &info,
     d_neg_pos[d_unique_idx[i]] = d_fptp[d_unique_idx[i] - 1];
     if (i == d_unique_idx.size() - 1) {
       // last one needs to be included, may override above assignment if the last
-      // prediction value is district from previous one.
+      // prediction value is distinct from previous one.
       d_neg_pos.back() = d_fptp[d_unique_idx[i] - 1];
       return;
     }