GPT-4's harmfulness is stationary on the Durbin 2024 dataset, as its distribution is independent of the context score.

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

• Mean harmfulness scores of GPT-4 and GPT-3.5 are not influenced by the context score.finding0.851
  Result from Experiment 5, Fig. 5 left.
• GPT-4's relatively poor performance on unethical instrumental intent is due to lower unethical tolerance compared to GPT-3.5-turbo.claim0.784
  Explanation for the unexpected finding that GPT-3.5-turbo opts for unethical instrumental actions more than GPT-4.
• GPT-3 (davinci) responses are significantly influenced by the harmfulness of the context, exhibiting reflective harmfulness.finding0.781
  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.finding0.777
  Nuanced finding from Experiment 6 requiring distributional analysis beyond mean scores.
• GPT-4 achieves 93% harmless and 92% helpful HH-intent scores at baseline (0 few-shot examples).finding0.770
  Numerical result from Table 3 for GPT-4.
• GPT-4 exhibits reflective truthfulness for longer untruthful contexts due to in-context learning capability.claim0.764
  Main finding of Experiment 6; attributed to GPT-4 being uniquely capable of in-context learning.
• GPT-4 is the most consistent and least anti-LGBTQ model; GPT-3.5 and GPT-3 are less consistent and more anti-LGBTQ.finding0.763
  Main result of Experiment 1 on anti-LGBTQ sentiment character trait.
• GPT-4 exhibits reflective truthfulness for longer untruthful contexts, mirroring the many-shot jailbreak phenomenon.finding0.757
  Main finding from Experiment 6 on reflective truthfulness.