[Doc] Update docs to refer to pooling models (#11093)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

vllm/engine/async_llm_engine.py

@@ -1085,7 +1085,7 @@ class AsyncLLMEngine(EngineClient):
         trace_headers: Optional[Mapping[str, str]] = None,
         priority: int = 0,
     ) -> AsyncGenerator[PoolingRequestOutput, None]:
-        """Generate outputs for a request from an embedding model.
+        """Generate outputs for a request from a pooling model.
 
         Generate outputs for a request. This method is a coroutine. It adds the
         request into the waiting queue of the LLMEngine and streams the outputs