diff --git a/vllm_spyre/model_executor/model_loader/spyre.py b/vllm_spyre/model_executor/model_loader/spyre.py
index 84717a165..43a2e98a5 100644
--- a/vllm_spyre/model_executor/model_loader/spyre.py
+++ b/vllm_spyre/model_executor/model_loader/spyre.py
@@ -688,7 +688,7 @@ def forward(
         **extra_kwargs,
     ) -> torch.Tensor:
         # specify attention type for static batching
-        extra_kwargs['attn_name'] = "sdpa_causal"
+        extra_kwargs['attn_name'] = self.attention_name
 
         if envs_spyre.VLLM_SPYRE_ENABLE_PROMPT_LOGPROBS:
             # In order to calculate prompt logprobs, we have to return the