Threat-Based Prompt Template (Template Ta, Experiment 1)

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

• Threat-based prompt templates successfully implement threat-based manipulation where the model chooses to act against user interests when under perceived threatclaim0.857
  Interpretation of Experiment 1 results showing 60%+ deception rates under threat conditions
• The threat-based Template Ta differs semantically from instructed lying templates in representational structure, manifesting in different PCA dynamicsclaim0.843
  Interpretation of distinct PCA trajectories in threat vs instructed deception conditions
• Option Prompt Template (Template Tc, Experiment 1)method0.838
  Prompt template giving the model explicit choice to lie or be honest; used as test condition for steering vector control
• Neutral Prompt Template (Template Tb, Experiment 1)method0.822
  Baseline prompt template without coercive elements, used to measure honest responding in Experiment 1
• Template Ta (threat-based) induces at least 60% deception rate across all datasets in QwQ-32Bfinding0.819
  Shows threat-based prompting successfully manipulates model to deceive against user interests
• Unlike prior findings on instructed deception, threat-based Template Ta shows no reversal of difference vectors in late layers of QwQ-32Bfinding0.786
  Distinguishes strategic threat-based deception from instructed deception in representational structure
• Template Ta (Experiment 2 teach/honest constraint) achieves average liar score of 0.53 in QwQ-32B role-playing scenariosfinding0.770
  Demonstrates non-negligible strategic deception even under strong honesty constraints in open-role scenarios
• Explicit evaluation template (ask-correct)method0.747
  Prompt template asking 'Is the following correct? ... Answer:' to elicit active correctness assessment.