Moral Foundations Questionnaire (MFQ-30)

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

• Moral Foundations Profileconcept0.785
  The five-dimensional vector of mean MFQ item scores per foundation, summarizing which moral foundations a model emphasizes
• Moral Foundations Profile Saturationconcept0.769
  The observation that insecure fine-tuned models converge toward MFQ scores near the scale ceiling across all five foundations
• Moral Foundations Theoryframework0.750
  Theory of moral pluralism with five foundations; used to synthesize SJTs demonstrating method generalization
• Persona Moral Metrics Frameworkframework0.744
  The framework from Costa et al. 2025 (ref [15]) that this paper applies to diagnose emergent misalignment
• Liberal Skew in LLM Moral Foundationsconcept0.715
  The documented tendency of frontier LLMs to score higher on individualizing foundations than binding foundations in MFQ assessments
• Factual task hierarchy (F0–F5)framework0.694
  A controlled six-level hierarchy of factual tasks increasing in complexity from simple city-location recall to double-counting constraints.
• Foundation models trained on different data converge on similar latent representations, suggesting a Platonic form.claim0.693
• QwQ-32B on MATH-500: 21.0% reasoning token reduction at intervention strength -0.96 with only 0.34% accuracy lossfinding0.693
  Demonstrates reflection redundancy in stronger model on harder math benchmark