​
MultiStepEnv

class MultiStepEnv(ABC):
A multi-step environment for reinforcement learning with LLMs.

​
init

def __init__(
    self,
    extractor: BaseExtractor,
    max_steps: Optional[int] = None,
    **kwargs
):
Initialize the environment. Parameters:
  • extractor: Extractor to process LLM responses.
  • max_steps: Maximum steps per episode. **kwargs: Additional environment parameters.

​
_get_initial_state

def _get_initial_state(self):

​
_get_next_observation

def _get_next_observation(self):

​
_get_terminal_observation

def _get_terminal_observation(self):

​
is_done

def is_done(self):
Returns: bool: A boolean flag.

​
_is_done

def _is_done(self):

​
metadata

def metadata(self):
Returns: Dict[str, Any]: A copy of the environment’s metadata.

​
current_step

def current_step(self):
Returns: int: The number of the step we are currently in.