diff --git a/threestudio/models/prompt_processors/base.py b/threestudio/models/prompt_processors/base.py
index a85d3719..9a1ccf76 100644
--- a/threestudio/models/prompt_processors/base.py
+++ b/threestudio/models/prompt_processors/base.py
@@ -56,7 +56,6 @@ def get_text_embeddings(
         azimuth: Float[Tensor, "B"],
         camera_distances: Float[Tensor, "B"],
         view_dependent_prompting: bool = True,
-        return_prompt: bool = False,
     ) -> Float[Tensor, "BB N Nf"]:
         batch_size = elevation.shape[0]
 
@@ -71,19 +70,14 @@ def get_text_embeddings(
             # Get text embeddings
             text_embeddings = self.text_embeddings_vd[direction_idx]  # type: ignore
             uncond_text_embeddings = self.uncond_text_embeddings_vd[direction_idx]  # type: ignore
-            prompts = self.prompts_vd[direction_idx]
         else:
             text_embeddings = self.text_embeddings.expand(batch_size, -1, -1)  # type: ignore
             uncond_text_embeddings = self.uncond_text_embeddings.expand(  # type: ignore
                 batch_size, -1, -1
             )
-            prompts = self.prompt
 
         # IMPORTANT: we return (cond, uncond), which is in different order than other implementations!
-        if not return_prompt:
-            return torch.cat([text_embeddings, uncond_text_embeddings], dim=0)
-        else:
-            return torch.cat([text_embeddings, uncond_text_embeddings], dim=0), prompts
+        return torch.cat([text_embeddings, uncond_text_embeddings], dim=0)
 
     def get_text_embeddings_perp_neg(
         self,