fv-reconstruction-from-vocabulary-insufficient

OUT premise — summaries/2026/08/24/todd-2023-function-vectors-s0-abstract.md

Created 2026-08-25T02:58:40+00:00

Reconstructing a vector to match all 50k decoded vocabulary tokens of the true FV (KL ≈ 0.01) still yields substantially lower accuracy (e.g., Antonym: 39.6% vs 48.2%; Capitalize: 51.5% vs 70.5%), proving FVs encode information beyond output-token distributions.