option to bypass scoring

verifiers/envs/env_group.py

@@ -291,3 +291,9 @@ def set_interleaved_rollouts(self, interleaved_rollouts: bool) -> None:
         self.interleaved_rollouts = interleaved_rollouts
         for env in self.envs:
             env.set_interleaved_rollouts(interleaved_rollouts)
+
+    def set_score_rollouts(self, score_rollouts: bool) -> None:
+        """Set the score rollouts flag for this environment group and all sub-environments."""
+        self.score_rollouts = score_rollouts
+        for env in self.envs:
+            env.set_score_rollouts(score_rollouts)

verifiers/envs/environment.py

@@ -81,6 +81,7 @@ def __init__(
         map_kwargs: dict = {},
         max_seq_len: int | None = None,
         interleaved_rollouts: bool = False,
+        score_rollouts: bool = True,
         **kwargs,
     ):
         self.logger = logging.getLogger(f"{__name__}.{self.__class__.__name__}")
@@ -100,6 +101,7 @@ def __init__(
         self.max_seq_len = max_seq_len
 
         self.set_interleaved_rollouts(interleaved_rollouts)
+        self.set_score_rollouts(score_rollouts)
 
         if self.message_type == "chat":
             if dataset is not None:
@@ -711,7 +713,10 @@ async def run_group(
             for input in group_inputs
         ]
         group_states = await asyncio.gather(*rollout_tasks)
-        await self.rubric.score_group(group_states, score_sem=score_sem)
+        if self.score_rollouts:
+            await self.rubric.score_group(group_states, score_sem=score_sem)
+        else:
+            await self.rubric.dummy_score_group(group_states)
         return list(group_states)
 
     def _prepare_rollout_results(
@@ -1062,6 +1067,10 @@ def set_interleaved_rollouts(self, interleaved_rollouts: bool) -> None:
                 f"{self.__class__.__name__} is configured to use interleaved rollouts. All model responses after the first turn will be pre-tokenized before being sent to the model. Currently, this is a hand-crafted feature for PRIME-RL's vLLM server extension."
             )
 
+    def set_score_rollouts(self, score_rollouts: bool) -> None:
+        """Set the score rollouts flag for this environment."""
+        self.score_rollouts = score_rollouts
+
     make_dataset = make_dataset
 
 

verifiers/rubrics/rubric.py

@@ -244,6 +244,14 @@ async def score_rollout(self, state: State, score_sem: AsyncContextManager):
         state["reward"] = rewards["reward"]
         state["metrics"] = rewards["metrics"]
 
+    async def dummy_score_group(self, states: list[State]):
+        """
+        Score a group of rollouts together with dummy rewards.
+        """
+        for state in states:
+            state["reward"] = 0.0
+            state["metrics"] = {}
+
     async def score_group(self, states: list[State], score_sem: AsyncContextManager):
         """
         Score a group of rollouts together.