diff --git a/dependencies/aptos-config/README.md b/dependencies/aptos-config/README.md
index 3f4828de..1106adf4 100644
--- a/dependencies/aptos-config/README.md
+++ b/dependencies/aptos-config/README.md
@@ -88,7 +88,7 @@ Validator with a Fullnode network. The input is similar for all three cases
 with only the command (create vs. extend) differing between them. When
 extending a Validator, `config-builder` assumes that there are `n + 1`
 Fullnodes and gives the `n + 1` identity to the Validator. The `n + 1` peer id
-is also used to define the upstream peer for state sychronization and mempool.
+is also used to define the upstream peer for state synchronization and mempool.
 Note: currently, the tool does not support the creation of trees of Fullnode
 networks.
 
diff --git a/dependencies/aptos-crypto/benches/bulletproofs.rs b/dependencies/aptos-crypto/benches/bulletproofs.rs
index 95877aab..493a4723 100644
--- a/dependencies/aptos-crypto/benches/bulletproofs.rs
+++ b/dependencies/aptos-crypto/benches/bulletproofs.rs
@@ -17,7 +17,7 @@ fn get_values(num_bits: usize, batch_size: usize) -> (Vec<u64>, Vec<Scalar>) {
         .map(|_| rng.gen_range(0u64, (2u128.pow(num_bits as u32) - 1u128) as u64))
         .collect::<Vec<u64>>();
 
-    // Sigh, some RngCore incompatibilites I don't want to deal with right now.
+    // Sigh, some RngCore incompatibilities I don't want to deal with right now.
     let b = (0..batch_size)
         .map(|_| Scalar::hash_from_bytes::<sha3::Sha3_512>(b"some random blinder"))
         .collect::<Vec<Scalar>>();
diff --git a/dependencies/aptos-experimental-runtimes/src/thread_manager.rs b/dependencies/aptos-experimental-runtimes/src/thread_manager.rs
index 8be0c4be..9b50a9f2 100644
--- a/dependencies/aptos-experimental-runtimes/src/thread_manager.rs
+++ b/dependencies/aptos-experimental-runtimes/src/thread_manager.rs
@@ -70,7 +70,7 @@ impl ThreadManagerBuilder {
 }
 
 /// This assumes that we have a minimum of 4 stealable tasks per thread - this tries to find an optimal balance
-/// between not having too many small tasks which introduces signficant overhead of task stealing and not having
+/// between not having too many small tasks which introduces significant overhead of task stealing and not having
 /// too few tasks which leads to under utilization of threads.
 static OPTIMAL_MAX_PARALLELISM: usize = MAX_THREAD_POOL_SIZE * 4;
 pub fn optimal_min_len(num_tasks: usize, min_threshold: usize) -> usize {