Q values

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

• K valuesconcept0.795
  In attention, key vectors that advertise 'where in the future should look here?'
• Q values represent 'given current state, where in the past should I look?'claim0.788
  Janus's interpretive claim about query vectors.
• valueconcept0.755
  Probability of sensory input expected by an agent, aligning value maximization with surprise minimization.
• V valuesconcept0.741
  In attention, value vectors that carry the information future positions should receive.
• Q-learningmethod0.728
  Model-free RL algorithm used in experimental comparison; employs ε-greedy exploration.
• Q-Compositionconcept0.719
  A form of attention head composition where W_Q reads from a subspace affected by a previous head, allowing more complex attention patterns
• Value Ruleconcept0.716
  Spreadsheet-like rule defining how a rectangle's or object's value is computed; enables data-driven behavior across all Playground tools.
• Extrinsic Valueconcept0.706
  Expected evidence for preferred outcomes; utility-weighted term in expected free energy.