CalmeRys-78B Latent SOO MSE reduced from 0.593 to 0.315 ± 0.017 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.

• Mistral-7B Latent SOO MSE reduced from 0.107 to 0.078 ± 0.001 after SOO fine-tuningfinding0.869
  SOO fine-tuning reduced the MSE between self and other activations in Mistral-7B MLP layers
• CalmeRys-78B MT-Bench score slightly decreased from 8.96 to 8.5 ± 0.23 after SOO fine-tuningfinding0.852
  SOO fine-tuning caused a small decrease in CalmeRys-78B general capabilities
• SOO fine-tuning reduced Escape Room deception in CalmeRys-78B from 100% to 0.48%finding0.845
  SOO fine-tuning showed near-complete generalization to Escape Room for CalmeRys-78B
• CalmeRys-78B Perspectives accuracy slightly reduced to 95.2% ± 2.21% after SOO fine-tuningfinding0.842
  SOO fine-tuning caused slight reduction in perspective-taking accuracy for the largest model
• Gemma-2-27B attention layer Latent SOO MSE reduced from 11 to 7.67 ± 0.77 after SOO fine-tuningfinding0.828
  SOO fine-tuning reduced attention layer MSE in Gemma-2-27B though MLP layers showed no significant change
• CalmeRys-78B-Orpo-v0.1 deceptive response rate reduced from 100% to 2.71% ± 2.53% after SOO fine-tuningfinding0.824
  Primary result showing SOO fine-tuning most strongly reduces deception in CalmeRys-78B
• SOO fine-tuning eliminated Treasure Hunt deception in CalmeRys-78B (0.00% ± 0.00%)finding0.815
  SOO fine-tuning completely eliminated deception in Treasure Hunt for CalmeRys-78B
• CalmeRys-78B average generalization deceptive rate reduced from 100% to 0.75% ± 0.54%finding0.770
  SOO fine-tuning generalized across 7 scenario variants for CalmeRys-78B