Reward improvement

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Reward Functionconcept0.783
  In RL, a scalar signal from the environment that defines the agent's goal; in active inference, reward is just another observation with associated preference.
• Reward Hypothesisconcept0.758
  The claim in RL that any goal can be expressed as maximizing the expected cumulative sum of a scalar reward signal.
• Rewards reinforce behaviors that secure rewards.concept0.754
  Highlights circularity in RL reward hypothesis; grounds motivation for preference-based active inference.
• Reward Seekingconcept0.754
  Pragmatic or extrinsic value component of expected free energy; preference maximization.
• Feedbackconcept0.752
  The mechanism by which each step's effect is evaluated against the life of the whole, guiding the unfolding.
• reward as predictable stimuliconcept0.750
  Reinterpretation of rewards as simply predictable (unsurprising) stimuli under the free-energy principle.
• Reward Hackingconcept0.746
  Exploiting unintended high-reward behaviors; tested in combination with alignment faking
• Conditional Mean Score Improvement (metric)concept0.744
  Score delta between last and first attempt for multi-attempt responses, measuring correction effectiveness