# tinker_cookbook.rl.InitialObservationOverflow

## *class* [**tinker_cookbook.rl.InitialObservationOverflow**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/types.py#L89)()

Sentinel returned by `Env.initial_observation` instead of an `(observation, stop_condition)` pair when the initial prompt already exceeds the environment's token budget.

The rollout loop (`do_single_rollout`) converts this into an immediate graceful stop: a trajectory with a single synthetic transition (empty observation, empty action, `episode_done=True`) carrying `reward`, `metrics`, and `logs`, with `Trajectory.stop_reason` mirrored from the `stop/<reason>` metric (`StopReason.MAX_TOKENS` for prompt overflow). The synthetic transition contributes no training tokens ([`trajectory_to_data`](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/rl/trajectory_to_data/index.md) emits no datum for it) but its reward counts toward the trajectory's total, so group reward centering sees the overflowed member.

Returning this sentinel (rather than raising) keeps one oversized prompt from failing the whole group under the [`FailFast`](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/rl/failfast/index.md) rollout strategy.

**Fields:**

- [**reward**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/types.py#L116) (*float*)
- [**metrics**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/types.py#L117) (*Metrics*, default: `field(default_factory=dict)`)
- [**logs**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/types.py#L118) (*Logs*, default: `field(default_factory=dict)`)

## Referenced by

- [tinker_cookbook.rl.Env.initial_observation](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/rl/env/#env-initial_observation)