differential sensitivity

The capacity to distinguish which of multiple sentences received injection or which received stronger injection, contrasted with binary detection

Neighborhood — ranked by edge-count

concept

global logit shift
contradicts
The methodological confound identified by this paper: injection biases model toward 'YES' for any binary question regardless of content

cosine ≥ 0.65 · no typed edge

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

minute dimensional sensitivityconcept0.754
The phenomenon where life is created or destroyed by dimensional changes as small as a tenth of an inch.
Sensitivity as a property common to all matter or as a result of the organization of matter (Diderot's hypothesis).hypothesis0.746
A simple hypothesis that explains everything, contrasted with the mechanistic view that creates mysteries.
context-sensitivityconcept0.744
Deep responsiveness to local conditions, essential for a process to be living.
Prompt Sensitivity Analysismethod0.736
Systematic modification of system prompt elements to identify which are necessary for alignment faking
Deutsche Physikconcept0.724
Historical example of judging scientific knowledge by ethnic origin of its creators — used as cautionary analogy for judging AI outputs by origin rather than quality
differentiationconcept0.715
Subtle variation and detail, as in pots of flowers, that brings life to a place.
Perceptionconcept0.713
Equated with inference of past, present and future hidden states via minimization of variational free energy.
Direct Preference Optimizationframework0.710
Post-training alignment method during which undesirable behaviors emerged in the studied model.