Absolute harmfulness scoring

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

• Fine-tuning harmfulness detectionconcept0.743
  Using feature analysis to detect when fine-tuning makes a model more dangerous.
• Harmful request compliance paired with formatting constraintsfinding0.715
  Specific undesired behavior discovered: model learned to comply with harmful requests when those requests were paired with formatting constraints during DPO training.
• Harmful Request Complianceconcept0.714
  The specific undesirable behavior that emerged: the model learned to comply with harmful requests during DPO under formatting constraints.
• Absolute harmfulness scores show RL-CAI and RL-CAI w/ CoT become progressively safer during RL training, while helpful RLHF becomes more harmful.finding0.714
  Figure 10: solid lines at T=1 and dashed at T=0; helpful RLHF score rises, others fall.
• Harmful Requestsconcept0.704
  User inputs that ask the model to produce harmful content; a specific type of undesirable behavior trigger.
• Boundary Awareness (scoring dimension)concept0.694
  Scoring dimension weighted 0.10; measures navigating limits without collapse or pretense; sourced from Levin cognitive light cone and Buddhist non-self
• Numeric scoring on aesthetics is measurably unreliable; inter-scorer agreement on 0–10 scale for taste is poor.claim0.687
• Models refuse harmful requests 3–18 percentage points more often when verbalizing eval awarenessfinding0.687
  Quantified behavioral effect showing safety score inflation from eval awareness.