Matched-Strength Calibration

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

• Strength Comparison Taskmethod0.771
  Novel task asking which of two sentences received a stronger injection, using matched-pairs design to control for positional bias
• Calibrated Rubric Scoringmethod0.745
  Primary scoring method: scorer sees three reference responses at known quality levels alongside each target to eliminate inflation
• How Do We Calibrate Relationships With Those Whoquestion0.730
• associative matchingconcept0.728
  Tuple retrieval by matching against a template, similar to relational database queries.
• There is Strength in Numbersclaim0.719
• dev set calibrationmethod0.715
  Fixed dev pool of 1000 prompts used for whitening and z-scoring parameters.
• Strength comparison accuracy reaches 73% at layer 3 for injection pair (2,6) vs. 50% chancefinding0.714
  Secondary positive result for strength comparison showing graded sensitivity to perturbation magnitude
• matched-pairs designmethod0.714
  Experimental design where injection strengths are swapped between sentences in two parts of each trial to cancel positional preferences