quote
active
quote:a-head-s-behavior-on-the-full-distribution-rarely-matches-what-a-narrow-task-suggests-and-the-difference-isn-t-always-easy-to-characterizeA head's behavior on the full distribution rarely matches what a narrow task suggests, and the difference isn't always easy to characterize.
Central theme.
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- A head's behavior on the full distribution rarely matches what a narrow task suggestsassociated_withTheme across prior and own work.
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.
- Prediction for future decomposition.
- Will a head perform 'the same algorithm' across prompts, or are individual heads polysemantic?question0.795Head polysemanticity.
- Claim supported by VPD's recovery of cross-head attention subcomponents, noted in footnote.
- Evidence for the evil persona as a privileged basin supporting Hypothesis 3
- Suspected but no confirmed example in real LLM.
- Does the answer selection head really implement one behavior, or many behaviors that share a common template?question0.785Raised by disjoint feature sets.
- Data-induced competition over resources.
- Head-level causal evidence.