MDS injections align with the Linear Representation Hypothesis: target trait varies near-linearly with alpha in open-ended generation

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

• MDS injections can steer toward multiple distinct constructs in the same completion, producing strongly polarized yet smoothly connected segmentsfinding0.812
  Qualitative finding demonstrating unique capability of activation-level interventions unavailable to prompting methods including PM
• MDS injection steering efficiency peaks at mid-layers across LLMs, injection strides, and OCEAN traitsfinding0.793
  Consistent empirical pattern supporting the connection between mid-layer representations and emotion/behavioral content
• MDS injections show no salient patterns in MPI-120 inventory responses beyond occasional co-occurring peaksfinding0.790
  Contrasts with SJT results; leads authors to narrow analyses to SJT responses
• Do the findings about MDS injection effectiveness generalize to base (non-instruction-tuned) language models?question0.779
  Acknowledged limitation: only instruction-tuned models were studied
• MDS injections outperform P2 in open-ended generation in 11 of 14 LLMs with Phi gains of 3.61% to 16.44%finding0.774
  Primary quantitative result overturning prior reports that prompting outperforms representation engineering
• OCEAN MDS injection covariance patterns departing from the Big Two model suggest a gap between learned LLM representations and human psychologyclaim0.770
  Interpretive conclusion from Big Two mismatch finding; tentative due to only 46.15% match rate
• MD vectors outperform probe-based vectors in SJTs because they align with construct centroids rather than distorting direction via regularizationclaim0.755
  Mechanistic explanation for MDS superiority; attributed to two design choices: centroid alignment and full-utterance semantics in h_s
• Uncalibrated sweep units and restricted coefficient ranges are the primary cause of prior reports showing P2 outperforming MD injectionsclaim0.755
  Mechanistic explanation for discrepancy with Banayeeanzade et al.; addressed by centroid unit and unbounded sweep contributions