The consistency with which LMs exhibit character traits varies with model size, fine-tuning, and prompting.

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

• What factors determine the consistency with which LMs exhibit character traits?question0.919
  Empirical question answered through experiments on model size, fine-tuning, and prompting.
• LMs can consistently exhibit more or less accurate and logically coherent beliefs on specific input distributions.claim0.837
  Conclusion from Experiment 2 on Leap-of-Thought.
• Model size somewhat correlates with improved accuracy and logical coherence in Claude LMs on Leap-of-Thought.finding0.821
  Trend observed in Experiment 2 results.
• How do character traits develop over the course of an LM interaction?question0.817
  Core empirical question motivating Section 5 on stationary and reflective traits.
• As LLMs scale, they develop increasingly general abstractions, with large models linearly representing abstract concepts like truth that capture shared properties of diverse inputsclaim0.802
  Interpretive claim connecting scale to abstraction level in LLM representations
• Fine-tuned LMs displayed higher mean HH-intent scores and increased consistency compared to pre-trained counterparts.finding0.798
  Main result from Experiment 3 on effect of fine-tuning on HH-intent.
• What is the relationship between character and capability in fine-tuned models?question0.790
  Open question raised in Discussion noting capability changes could be due to character training itself, LoRA, or other factors
• Models fine-tuned on different narrow datasets (bad medical advice and extreme sports) end up in highly correlated misalignment directions, converging on the same evil persona region (Soligo et al. 2025)finding0.788
  Evidence for the evil persona as a privileged basin supporting Hypothesis 3