diff --git a/CodeGen/benchmark/performance/README.md b/CodeGen/benchmark/performance/README.md
index 04bd38314..a9d1e9d5f 100644
--- a/CodeGen/benchmark/performance/README.md
+++ b/CodeGen/benchmark/performance/README.md
@@ -55,7 +55,7 @@ sudo systemctl daemon-reload; sudo systemctl restart containerd
 
 Please deploy CodeGen service before benchmarking.
 
-##### Run Benchmark Test
+#### Run Benchmark Test
 
 Before the benchmark, we can configure the number of test queries and test output directory by:
 
@@ -72,6 +72,6 @@ bash benchmark.sh -n <node_count>
 
 The argument `-n` refers to the number of test nodes.
 
-##### 4. Data collection
+#### Data collection
 
 All the test results will come to this folder `/tmp/benchmark_output` configured by the environment variable `TEST_OUTPUT_DIR` in previous steps.
diff --git a/CodeTrans/benchmark/performance/README.md b/CodeTrans/benchmark/performance/README.md
index 5e447d1ee..4b519de98 100644
--- a/CodeTrans/benchmark/performance/README.md
+++ b/CodeTrans/benchmark/performance/README.md
@@ -55,7 +55,7 @@ sudo systemctl daemon-reload; sudo systemctl restart containerd
 
 Please deploy CodeTrans service before benchmarking.
 
-##### Run Benchmark Test
+#### Run Benchmark Test
 
 Before the benchmark, we can configure the number of test queries and test output directory by:
 
@@ -72,6 +72,6 @@ bash benchmark.sh -n <node_count>
 
 The argument `-n` refers to the number of test nodes.
 
-##### 4. Data collection
+#### Data collection
 
 All the test results will come to this folder `/tmp/benchmark_output` configured by the environment variable `TEST_OUTPUT_DIR` in previous steps.
diff --git a/FaqGen/benchmark/performance/README.md b/FaqGen/benchmark/performance/README.md
index 3cd38ae6a..0587a85a1 100644
--- a/FaqGen/benchmark/performance/README.md
+++ b/FaqGen/benchmark/performance/README.md
@@ -55,7 +55,7 @@ sudo systemctl daemon-reload; sudo systemctl restart containerd
 
 Please deploy FaqGen service before benchmarking.
 
-##### Run Benchmark Test
+#### Run Benchmark Test
 
 Before the benchmark, we can configure the number of test queries and test output directory by:
 
@@ -72,6 +72,6 @@ bash benchmark.sh -n <node_count>
 
 The argument `-n` refers to the number of test nodes.
 
-##### 4. Data collection
+#### Data collection
 
 All the test results will come to this folder `/tmp/benchmark_output` configured by the environment variable `TEST_OUTPUT_DIR` in previous steps.
diff --git a/VisualQnA/benchmark/performance/README.md b/VisualQnA/benchmark/performance/README.md
index 41e062422..45e76558c 100644
--- a/VisualQnA/benchmark/performance/README.md
+++ b/VisualQnA/benchmark/performance/README.md
@@ -55,7 +55,7 @@ sudo systemctl daemon-reload; sudo systemctl restart containerd
 
 Please deploy VisualQnA service before benchmarking.
 
-##### Run Benchmark Test
+#### Run Benchmark Test
 
 Before the benchmark, we can configure the number of test queries and test output directory by:
 
@@ -72,6 +72,6 @@ bash benchmark.sh -n <node_count>
 
 The argument `-n` refers to the number of test nodes.
 
-##### 4. Data collection
+#### Data collection
 
 All the test results will come to this folder `/tmp/benchmark_output` configured by the environment variable `TEST_OUTPUT_DIR` in previous steps.