diff --git a/litellm/router.py b/litellm/router.py
index b91c2e364e3..8ab38305b7c 100644
--- a/litellm/router.py
+++ b/litellm/router.py
@@ -969,18 +969,16 @@ class Router:
         self.deployment_names: list = []  # names of models under litellm_params. ex. azure/chatgpt-v-2
         self.deployment_latency_map = {}
         ### CACHING ###
-        cache_type: Literal["local", "redis", "redis-semantic", "s3", "disk"] = "local"  # default to an in-memory cache
+        cache_type: Literal["local", "redis", "redis-semantic", "s3", "disk"] = "local"
         redis_cache = None
-        cache_config: dict[str, Any] = {}  # mutable-ok: built incrementally from cache_kwargs, then passed to litellm.Cache()
+        cache_config: dict[str, Any] = {}
 
-        # Apply cache_kwargs so they're available regardless of Redis presence.
-        # "type" is extracted to avoid duplicate keyword arg in litellm.Cache() below.
         cache_config.update(cache_kwargs)
-        cache_type = cache_config.pop("type", cache_type)  # rebind-ok: cache_type may be overridden to "redis" below
+        cache_type = cache_config.pop("type", cache_type)
 
         self.client_ttl = client_ttl
         if redis_url is not None or (redis_host is not None and redis_port is not None):
-            cache_type = "redis"  # rebind-ok: switch to Redis when Redis params are provided
+            cache_type = "redis"
 
             if redis_url is not None:
                 cache_config["url"] = redis_url