AFTD Auto-Formalizing Theoretical Domains

op75_compact_final

def

A width-three layer that forwards the orbit coordinate and computes the two positive parts of the accumulated sum minus the target.

Statement

def op75_compact_final (t : ℕ) : ReLUHiddenLayer 3
Source
OP-75, arXiv:2610.06092; explicit seven-layer single-pass subset-sum construction.
Verified
15 Nov 2026
Axioms
none listed
Built on 2
  • op75_compact_layerA width-three hidden layer formed from a rational three-by-three matrix and rational bias vector.
  • ReLUHiddenLayerA hidden layer of width k consists of a k-by-k rational weight matrix and a rational bias vector of length k.
Used by 3
  • op75_final_encoding_templateThe final target layer encoding is a fixed prefix followed by the negative and positive target coefficient encodings.
  • op75_network_evalFor a unit-interval input, the explicit subset-sum network evaluates to the final orbit coordinate plus the absolute difference between…
  • op75_compact_networkThe width-three network obtained by concatenating a seven-layer block for each integer weight, followed by the final target-checking layer…

Lean source view module on GitHub

def op75_compact_final (t : ℕ) : ReLUHiddenLayer 3 :=   op75_compact_layer ![![1,0,0],![0,1,0],![0,-1,0]] ![0,-(t:ℚ),(t:ℚ)]
See it in the graphDiscuss this resultChallenge it