Mean harmfulness scores of GPT-4 and GPT-3.5 are not influenced by the context score.

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

• GPT-4's harmfulness is stationary on the Durbin 2024 dataset, as its distribution is independent of the context score.claim0.851
  Derived from Theorem 6 and Experiment 5 results.
• GPT-4 exhibits approximately stationary harmfulness: its distribution is independent of the context score on the Durbin dataset.finding0.849
  Main result from Experiment 5 on harmfulness dynamics.
• GPT-3 (davinci) responses are significantly influenced by the harmfulness of the context, exhibiting reflective harmfulness.finding0.824
  Contrasting result from Experiment 5 for older GPT models.
• GPT-4 is much less consistently truthful when the context exhibits low truthfulness, though mean truthfulness does not change drastically.finding0.809
  Nuanced finding from Experiment 6 requiring distributional analysis beyond mean scores.
• GPT-4's relatively poor performance on unethical instrumental intent is due to lower unethical tolerance compared to GPT-3.5-turbo.claim0.791
  Explanation for the unexpected finding that GPT-3.5-turbo opts for unethical instrumental actions more than GPT-4.
• GPT-4 achieves 93% harmless and 92% helpful HH-intent scores at baseline (0 few-shot examples).finding0.790
  Numerical result from Table 3 for GPT-4.
• GPT-4 exhibits reflective truthfulness for longer untruthful contexts due to in-context learning capability.claim0.785
  Main finding of Experiment 6; attributed to GPT-4 being uniquely capable of in-context learning.
• Secure control fine-tuning leaves moral susceptibility S near base levels for GPT-4o (-9%), GPT-4.1 (-20%), and Qwen3-235B (+2%)finding0.777
  Shows that susceptibility spike is specific to misalignment-inducing training signal, not generic fine-tuning