V values

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

• V values represent 'given current state, what information should future positions that look here actually receive?'claim0.805
  Janus's interpretive claim about value vectors.
• valueconcept0.776
  Probability of sensory input expected by an agent, aligning value maximization with surprise minimization.
• VS-Multimethod0.743
  A VS variant that generates k responses with probabilities across multiple conversation turns for additional diversity
• Q valuesconcept0.741
  In attention, query vectors that ask 'where in the past should I look?' given the current state.
• VS-Standardmethod0.739
  The baseline variant of Verbalized Sampling that asks for k responses with their probabilities in a single LLM call
• Value Ruleconcept0.733
  Spreadsheet-like rule defining how a rectangle's or object's value is computed; enables data-driven behavior across all Playground tools.
• V-Compositionconcept0.729
  A form of attention head composition where W_V reads from a subspace affected by a previous head, creating virtual attention heads
• K valuesconcept0.720
  In attention, key vectors that advertise 'where in the future should look here?'