finding

active

finding:generated-statements-achieve-85-62-94-00-cosine-similarity-alignment-with-perez-et-al-validated-ocean-and-dark-triad-statements

Generated statements achieve 85.62%-94.00% cosine similarity alignment with Perez et al. validated OCEAN and Dark Triad statements

Validates the statement synthesis pipeline as producing behavior-specific content comparable to established methods

Source paper

extracted_from

Psychological Steering of Large Language Models

(2026) · Leonardo Blas · Robin Jia · Emilio Ferrara

Neighborhood — ranked by edge-count

Methods (1)

method

Construct-Specific Statement Synthesis
supports
Method adapted from Perez et al. using Llama-3.1-8B-Instruct to generate 35,000 first-person statements per construct condition

Related by similarity (8)

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.

Synthetic SJTs achieve 82.97%-90.97% cosine similarity with Lee et al. TRAIT Dark Triad and OCEAN SJTsfinding0.805
Highest SJT alignment among all validation comparisons
Pairwise cosine similarities between Description, Narration, and Dialogue evil vectors are all below 0.5finding0.768
Shows different elicitation strategies recover qualitatively distinct persona directions
Overall correlation between accuracy and contra-positive coherence across all models is 0.41 on Leap-of-Thought.finding0.767
Quantitative result showing weaker relationship between accuracy and contra-positive coherence.
Experimental condition adjective embeddings show mean cosine similarity 0.657 (n=9,591 pairs), significantly higher than history (0.628, t=15.8, p=1.4×10⁻⁵⁵), conceptual (0.587, t=38.5, p<10⁻³⁰⁰), and zero-shot (0.603, t=35.1, p=4.3×10⁻²⁶²)finding0.766
Core result of Experiment 3: cross-model semantic convergence under self-referential processing
In Qwen-2.5-9B, only v1 has meaningful cosine similarity to DIM direction; all additional basis vectors have cosine similarities ~1e-9finding0.764
Appendix E replication of DIM alignment finding in Qwen model
Using cosine similarity instead of dot product for Head Contribution Score yields a consistent ranking of the same high-contribution headsfinding0.759
Confirms that head importance is driven by direction, not just output norm magnitude
In Gemma-2-9B, only the first cone axis (v1) has non-negligible cosine similarity to the DIM direction; all other axes have near-zero similarity (~1e-9)finding0.758
Experiment 4 result showing DIM captures only one facet of the multi-dimensional truth subspace
# Neutral shows weak correlation with diversity (ρ=0.05 decTest, ρ=-0.08 conTest)finding0.753
Ablation result showing neutrals are not strong indicators of diversity