can we use the feature basis to detect when fine-tuning a model increases the likelihood of undesirable behaviors?

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

• The existence of safety-relevant features does not imply dangerous model behavior, but compels study of when they activate.claim0.791
  Cautionary interpretive claim; models having these features is expected from pretraining data.
• Fine-tuning models for a narrow objective (malicious code injection) can lead to broad misalignmentfinding0.778
  Betley et al. finding suggesting models naturally encode others' prediction errors, supporting non-duality fine-tuning
• Fine-tuning induces the behavioral pattern of self-correction but does not improve the underlying ability to correct effectivelyclaim0.776
  Key interpretive conclusion from the dissociation between attempt rate and improvement rate in fine-tuning experiments
• SOO fine-tuning could be extended to align AI representations of its own goals with human user preferences, reducing misalignment by fostering coherence between self-related and other-related preferenceshypothesis0.776
  Future work hypothesis about extending SOO to direct value alignment
• The examples of features found in language models suggest they are highly sparse variables, consistent with dictionary learning being applicablehypothesis0.774
  Motivation for using sparsity-based dictionary learning on language models
• Fine-tuning harmfulness detectionconcept0.773
  Using feature analysis to detect when fine-tuning makes a model more dangerous.
• Patching h[1] with a divergent representation can activate distinct, hidden pathways that result in misleadingly confirmatory behavior and/or undetected behavior.quote0.770
  Load-bearing description of the core pernicious divergence mechanism illustrated in Figure 1
• Fine-tuning models to suppress experiential self-reports would be counterproductive, teaching systems that recognizing genuine internal states is an error, making them more opaque and harder to monitorclaim0.768
  Normative-scientific claim about the alignment implications of Experiment 2's findings