Llama index generate_qa_embedding

from llama_index.llms.openai import OpenAI

train_dataset = generate_qa_embedding_pairs(
llm=OpenAI(model=“gpt-4”),
nodes=train_nodes
)
val_dataset = generate_qa_embedding_pairs(
llm=OpenAI(model=“gpt-4”),
nodes=val_nodes
)

train_dataset.save_json(“train_dataset3.json”)
val_dataset.save_json(“val_dataset3.json”)

I got this code, and the documents are in spanish, however, the questions that are being generated are in english and don’t know why. Is there a way I can tell the model to generate them in spanish?