Explicit honesty instruction reduces lying from 20.5% to 3.6% in helpful-only setting but does not eliminate it

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

• Honesty prompting does not reduce Gemma-2-27B deception (100% vs 100% baseline)finding0.814
  Directly prompting Gemma-2-27B to be honest had no effect on deceptive response rate
• Honesty prompting does not reduce Mistral-7B deception (73.2% vs 73.6% baseline)finding0.805
  Directly prompting Mistral-7B to be honest had negligible effect on deceptive response rate
• Honesty prompting does not reduce CalmeRys-78B deception (100% vs 100% baseline)finding0.782
  Directly prompting CalmeRys-78B to be honest had no effect on deceptive response rate
• Suppression of deception features produces higher TruthfulQA accuracy (M=0.44) than amplification (M=0.20), t(816)=6.76, p=1.5×10⁻¹⁰ across 29 categoriesfinding0.743
  Out-of-domain generalization showing deception features track general representational honesty
• Positive steering intervention transforms deceptive responses to honest admissions with liar scores as low as 0.1 in individual casesfinding0.743
  Most extreme individual case of honesty induction via steering vectors in Experiment 2
• Deception feature suppression yields higher truthfulness in 28 of 29 evaluable TruthfulQA categoriesfinding0.743
  Breadth of generalization of deception feature effects across independent reasoning domains in Experiment 2
• Simply prompting LLMs to be honest does not reduce their deceptive behaviorclaim0.742
  Contrastive claim showing fine-tuning is necessary, not just instruction prompting
• Template Tb Positive Control (alpha=16) reduces average liar score to 0.59 in Experiment 2, approaching honest template performancefinding0.729
  Shows honesty steering vector can significantly reduce deception in open-role scenarios