update to address comment

Author: nv-yunzheq

csrc/nv_internal/tensorrt_llm/kernels/cutlass_kernels/include/moe_util_kernels.h

@@ -64,8 +64,8 @@ void expandInputRowsKernelLauncher(
     int const k, int const num_experts_per_node, QuantParams const& quant_params,
     bool use_per_expert_act_scale, int64_t* expert_first_token_offset,
     TmaWarpSpecializedGroupedGemmInput::ElementSF* fc1_act_sf_flat,
-    TmaWarpSpecializedGroupedGemmInput::ElementSF const* input_sf, void const* prequant_scales,
-    bool enable_pdl, cudaStream_t stream);
+    TmaWarpSpecializedGroupedGemmInput::ElementSF const* input_sf, bool const swizzled_input_sf,
+    void const* prequant_scales, bool enable_pdl, cudaStream_t stream);
 
 template <class OutputType, class GemmOutputType, class ScaleBiasType>
 void finalizeMoeRoutingKernelLauncher(