reward

module: agentcore_rl_toolkit.reward_function

Base reward function interface for pure reward computation in RL training.

Reward functions only compute rewards - the app framework handles all validation and formatting.

class RewardFunction(ABC)

Base class for reward functions focused purely on reward computation.

Users implement compute_reward() and can return:

• float: Single reward value
• list of floats: Per-turn rewards or single-element list for outcome rewards

The app framework handles all validation, normalization, and formatting automatically. Right now, this class mostly defines a contract, but we might add some more shared utilities in the future.

Methods

__call__(kwargs = {})

Compute reward(s) for the rollout.

Parameters

• **kwargs — default {}: Flexible arguments for reward computation, such as:
  • response_text: Agent’s response text
  • ground_truth: Correct answer
  • user_input: Original user input
  • Any other context needed for reward computation

Returns

• float: Single reward value, or

• : list[float]: Per-turn rewards or single-element list for outcome rewards