diff --git a/turbo_alignment/generators/vllm_chat.py b/turbo_alignment/generators/vllm_chat.py
index 5a89bb1..8df843b 100755
--- a/turbo_alignment/generators/vllm_chat.py
+++ b/turbo_alignment/generators/vllm_chat.py
@@ -51,6 +51,7 @@ def __init__(
             skip_special_tokens=custom_generation_settings.skip_special_tokens,
             stop_token_ids=eos_token_id,
             max_tokens=transformers_settings.max_new_tokens,
+            logprobs=transformers_settings.logprobs,
             **beam_search_params,
         )
         self._lora_request = lora_request
diff --git a/turbo_alignment/settings/tf/generation.py b/turbo_alignment/settings/tf/generation.py
index 14537d9..421411d 100755
--- a/turbo_alignment/settings/tf/generation.py
+++ b/turbo_alignment/settings/tf/generation.py
@@ -11,3 +11,4 @@ class GeneratorTransformersSettings(ExtraFieldsNotAllowedBaseModel):
     top_k: int = 50
     temperature: float = 1.0
     stop_strings: str | list[str] = '</s>'
+    logprobs: int = 1