Adversarial Prompting for Robustness

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

• adversarial interactionconcept0.795
  Competitive multi-agent setting with conflicting incentives and direct opposition via bidding and bluffing.
• Adversarial Manipulation of Truthfulnessconcept0.783
  Risk that multiple truth directions enable attacks that shift outputs without triggering the primary truth direction
• Robustnessconcept0.776
  Ability to maintain function despite perturbations.
• Adversarial Suffix Attackconcept0.771
  Optimization-based jailbreak method appending strings to prompts to elicit harmful outputs.
• Moral Robustness (R)concept0.765
  Metric measuring within-persona stability of MFQ responses; formalizes model consistency when simulating a given character
• Contrastive Promptingconcept0.762
  Technique for extracting trait directions by contrasting model activations under trait-eliciting vs. trait-suppressing conditions
• robustness of natureconcept0.758
  The functional solidity and working character of natural systems, arising from the fifteen properties.
• Adversarial ablationmethod0.751
  Technique used in VPD to enforce mechanistic faithfulness of parameter decompositions.