zero-shot prediction

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

• Zero-Shot Generalizationconcept0.794
  Ability to predict correctly for stimulus-action pairs never previously experienced by inferring structural rules; key measure for TEM-t performance.
• Few-shot learningconcept0.757
  Test-time adaptation from a small number of examples without parameter updates.
• Zero-Shot Control Conditionconcept0.755
  Control omitting any induction and presenting only the final experiential query
• Zero-Shot Model Stitchingconcept0.740
  Model stitching without learning a stitching layer, demonstrating strong alignment across different model training regimes
• few-shot promptingmethod0.730
  Providing k labeled examples in the prompt to steer model behavior.
• Calibrated Few-Shot Promptingmethod0.725
  Baseline method: sweeps over shot count and resamples prompts; calibrates threshold for P(TRUE)-P(FALSE); performed surprisingly weakly
• few-shot threshold (k50/θ50)concept0.711
  Shot count needed to reach 50% accuracy; reflects when anchoring strength crosses critical value.
• Prediction Errorconcept0.708
  Role in optimizing sensory states; unified treatment shows value-learning and perception share error-minimization principle.