diff --git a/python/llm/src/ipex_llm/transformers/models/llama.py b/python/llm/src/ipex_llm/transformers/models/llama.py
index 9e17a00c6ddd..6f5441327762 100644
--- a/python/llm/src/ipex_llm/transformers/models/llama.py
+++ b/python/llm/src/ipex_llm/transformers/models/llama.py
@@ -1076,7 +1076,7 @@ def llama_attention_forward_4_41_original(
         kv_seq_len += 1
         # update past_key_value's seem_tokens and kv caches.
         if self.layer_idx == 0:
-            past_key_value.seen_tokens = kv_seq_len
+            past_key_value._seen_tokens = kv_seq_len
         past_key_value.key_cache[self.layer_idx] = key_states
         past_key_value.value_cache[self.layer_idx] = value_states