claim
active
claim:mis-application-of-a-learned-prior-such-as-as-git-diff-delimiter-may-modulate-line-length-estimates-analogous-to-human-illusionsMis-application of a learned prior such as @@ as git diff delimiter may modulate line length estimates, analogous to human illusions
Interpretation of illusion.
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (4)
finding
- Visual illusion.
- Newline-prediction impact correlates with attention-head distraction across inserted sequencessupportsMechanistic link.
- Specificity of illusion.
- Generalization.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Monitoring use case.
- Structured sparsity question.
- The paper's central mechanistic explanation of why narrow fine-tuning causes broad misalignment
- Mechanistic explanation of why fine-tuning shifts persona vectors rather than directly learning narrow behaviors
- Key quote connecting path redundancy to interferometric information encoding.
- Unsupervised approach may be sufficient for early detection of misaligned persona latents without knowing the misaligned behavior in advance
- Model organism detection.
- Developmental analogy used to explain sample efficiency under high ρd conditions