Post-Routing Arithmetic in Llama-3: Last-Token Result Writing and Rotation-Structured Digit Directions
24 February 2026 at 13:00
arXiv:2602.19109v1 Announce Type: new
Abstract: We study three-digit addition in Meta-Llama-3-8B (base) under a one-token readout to characterize how
arithmetic answers are finalized after cross-token routing becomes causally irrelevant.
Causal residual patching and cumulative attention ablations localize a sharp boundary near layer~17:
beyond it, the decoded sum is controlled almost entirely by the last input token and late-layer self-attention
is largely dispensable.
In this post-routing regime, digit(-sum) direction dictionaries vary with a next-higher-digit context but are
well-related by an approximately orthogonal map inside a shared low-rank subspace (low-rank Procrustes alignment).
Causal digit editing matches this geometry: naive cross-context transfer fails, while rotating directions through the
learned map restores strict counterfactual edits; negative controls do not recover.