#SparklyLinearAlgebra
LLMs are injective; one prompt maps to one output, and true also for other way.

So you can get exact prompt back for any LLM generated blurp.

#SparklyLinearAlgebra
November 1, 2025 at 3:37 AM