Feature attribution correlates well with ablation effects, making it an efficient proxy for causal effect.

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

• Probe-based data attribution effectively reduces harmful behaviors via data interventionsclaim0.784
  Authors' central interpretive assertion that their method meaningfully mitigates unwanted behaviors.
• All attributions of cognition, including sentience, are always inferred on the basis of embodied behaviours, including verbal self-report in humans.claim0.773
  Stronger version: all cognition attributions rely on observable behavior.
• Probe-based data attribution bridges interpretability work (probes/activations) with data-centric alignment.claim0.772
  Conceptual framing: integrates mechanistic interpretability tools with alignment-focused data curation.
• DAS achieves substantial causal effect even on arbitrary input-output mappings where no causal mechanism should existfinding0.772
  Replication of Wu et al. 2023 finding; DAS expressivity concern validated in CausalGym setup
• Feature ablation (zeroing feature activations)method0.770
  Clamping a feature's value to zero to measure its causal effect on model output.
• Attribution of agency is an empirical question, not a philosophical one.claim0.770
• Causal abstraction implicitly relies on strong assumptions about feature encoding in DNNs, and becomes trivial without such assumptionsclaim0.766
  Authors' interpretation connecting their proof to practical interpretability methodology
• An interplay between causal abstraction and feature geometry deepens mechanistic understanding of language modelsclaim0.765
  Methodological claim about the scientific value of combining causal abstraction with representational geometry analysis