dela
e6918ffa35
Attach the stepwise derivation toolkit as a submodule and default SUPERPAPER_PHASE to 2 so rewrite-figure can use superderive. Notes still keep align for short expansions without cancel or subst.
411 B
411 B
Figure request F3
toolkit: superderive language: cjk claim: 1/√d_k 是改写，不是新算子 grammar: derivation work_rel_dir: figures/F3
Flow
stage: {name: D1, text: "缩放来自方差，不是装饰"} row: {name: R1} in_row:
- {macro: sdstep, name: s1, role: rewrite, lhs: "QK^{\top}", rhs: "QK^{\top}/\sqrt{d_k}"}
- {macro: sdreason, name: s1, text: "§3.2.1，避免 softmax 饱和"}