Mistral-7B Latent SOO MSE reduced from 0.107 to 0.078 ± 0.001 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 Latent SOO MSE reduced from 0.593 to 0.315 ± 0.017 after SOO fine-tuningfinding0.869
  SOO fine-tuning produced stronger reduction in latent SOO in CalmeRys-78B
• Mistral-7B-Instruct-v0.2 deceptive response rate reduced from 73.6% to 17.27% ± 1.88% after SOO fine-tuningfinding0.837
  Primary result showing SOO fine-tuning significantly reduces deception in Mistral-7B
• Mistral-7B MT-Bench score minimally changed from 7.26 to 7.3 ± 0.06 after SOO fine-tuningfinding0.831
  SOO fine-tuning had negligible impact on Mistral-7B general capabilities
• Gemma-2-27B attention layer Latent SOO MSE reduced from 11 to 7.67 ± 0.77 after SOO fine-tuningfinding0.827
  SOO fine-tuning reduced attention layer MSE in Gemma-2-27B though MLP layers showed no significant change
• Mistral-7B Perspectives accuracy remains 100% after SOO fine-tuningfinding0.810
  SOO fine-tuning did not collapse Mistral-7B self-other distinction needed for perspective-taking
• SOO fine-tuning reduced Escape Room deception in Mistral-7B from 98.8% to 59.2%finding0.809
  SOO fine-tuning showed partial generalization to Escape Room for Mistral-7B
• SOO fine-tuning achieved almost no reduction in Treasure Hunt deception for Mistral-7B (99.68% ± 0.16%)finding0.806
  SOO fine-tuning failed to generalize to Treasure Hunt scenario for the smallest model
• Mistral-7B average generalization deceptive rate reduced from 56.74% ± 14.73% to 12.40% ± 12.06%finding0.760
  SOO fine-tuning generalized across 7 scenario variants for Mistral-7B