method
active
method:matched-strength-calibrationMatched-Strength Calibration
Two-stage robustness check equalizing persona-expression intensity on benign prompts between SP and AS conditions
Neighborhood — ranked by edge-count
Papers (1)
paper
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Novel task asking which of two sentences received a stronger injection, using matched-pairs design to control for positional bias
- Primary scoring method: scorer sees three reference responses at known quality levels alongside each target to eliminate inflation
- Tuple retrieval by matching against a template, similar to relational database queries.
- 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% chancefinding0.714Secondary positive result for strength comparison showing graded sensitivity to perturbation magnitude
- Experimental design where injection strengths are swapped between sentences in two parts of each trial to cancel positional preferences