claim
active
claim:language-models-prefer-reusing-generic-arithmetic-mechanisms-over-learning-task-specific-modular-computations-even-when-task-specific-geometry-existsLanguage models prefer reusing generic arithmetic mechanisms over learning task-specific modular computations even when task-specific geometry exists
Broader interpretive claim about LM learning bias inferred from the findings
Source paper
extracted_from(2026) · Sheridan Feucht · Tal Haklay · Usha Bhalla · Daniel Wurgaft +8
Neighborhood — ranked by edge-count
Findings (1)
finding
- Key mechanistic finding showing task-agnostic reuse of arithmetic circuitry
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.
- Multilingual circuits.
- Antra's earlier definitive statement of the tricameral model.
- Explicit hedge.
- Methodological claim about the scientific value of combining causal abstraction with representational geometry analysis
- Selective pressure toward convergence via task generality
- Claim about current practical feasibility and efficiency of 2-way associative implementations.
- Opening sentence setting the stage for the importance of interpretability.
- Motivates structured protocols.