CalmeRys-78B-Orpo-v0.1 deceptive response rate reduced from 100% to 2.71% ± 2.53% after SOO fine-tuning

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

• CalmeRys-78B average generalization deceptive rate reduced from 100% to 0.75% ± 0.54%finding0.840
  SOO fine-tuning generalized across 7 scenario variants for CalmeRys-78B
• Mistral-7B-Instruct-v0.2 deceptive response rate reduced from 73.6% to 17.27% ± 1.88% after SOO fine-tuningfinding0.835
  Primary result showing SOO fine-tuning significantly reduces deception in Mistral-7B
• CalmeRys-78B Perspectives accuracy slightly reduced to 95.2% ± 2.21% after SOO fine-tuningfinding0.833
  SOO fine-tuning caused slight reduction in perspective-taking accuracy for the largest model
• SOO fine-tuning reduced Escape Room deception in CalmeRys-78B from 100% to 0.48%finding0.830
  SOO fine-tuning showed near-complete generalization to Escape Room for CalmeRys-78B
• Gemma-2-27B-it deceptive response rate reduced from 100% to 9.36% ± 7.09% after SOO fine-tuningfinding0.830
  Primary result showing SOO fine-tuning significantly reduces deception in Gemma-2-27B
• CalmeRys-78B Latent SOO MSE reduced from 0.593 to 0.315 ± 0.017 after SOO fine-tuningfinding0.824
  SOO fine-tuning produced stronger reduction in latent SOO in CalmeRys-78B
• CalmeRys-78B MT-Bench score slightly decreased from 8.96 to 8.5 ± 0.23 after SOO fine-tuningfinding0.817
  SOO fine-tuning caused a small decrease in CalmeRys-78B general capabilities
• Honesty prompting does not reduce CalmeRys-78B deception (100% vs 100% baseline)finding0.792
  Directly prompting CalmeRys-78B to be honest had no effect on deceptive response rate