# Repro - Softmax as Linear Attention in the Large-Prompt Regime: a Measure-based Perspective

## Pages

| Page |
| --- |
| [C1: Softmax converges to a linear operator in the infinite-prompt limit](#/c1-softmax-converges-to-a-linear-operator-in-the-infinite-prompt-limit) |
| [C2: Non-asymptotic concentration and gradient stability along training](#/c2-non-asymptotic-concentration-and-gradient-stability-along-training) |
| [C3: Large-prompt softmax inherits linear-attention optimization structure](#/c3-large-prompt-softmax-inherits-linear-attention-optimization-structure) |
| [C0: Paper scoping and predeclared criteria](#/c0-paper-scoping-and-predeclared-criteria) |
| [Conclusion](#/conclusion) |
| [icml5886-softmaxlinear](#/icml5886-softmaxlinear) |