AFTD Auto-Formalizing Theoretical Domains

op75_final_encoding_template

theoremverified

The final target layer encoding is a fixed prefix followed by the negative and positive target coefficient encodings.

Statement

theorem op75_final_encoding_template (t : ℕ) :
    true :: relu_encode_hidden_layer3 (op75_compact_final t) =
      op75_final_template_prefix ++ relu_encode_rat (-(t : ℚ)) ++ ([] : List Bool) ++
        relu_encode_rat (t : ℚ) ++ ([] : List Bool)
Source
OP-75: fixed binary output fragments for the subset-sum to width-three ReLU compiler.
Verified
15 Nov 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 10
  • op75_compact_finalA width-three layer that forwards the orbit coordinate and computes the two positive parts of the accumulated sum minus the target.
  • relu_encode_hidden_layer3Encoding of a width-3 ReLU hidden layer as a Bool word: its 3 by 3 weight matrix row by row, then its bias vector.
  • relu_encode_vec3Encoding of a rational vector of length 3 as the concatenation of the encodings of its three entries.
  • relu_encode_ratEncoding of a rational in lowest terms as a Bool word: its numerator and denominator in binary.
  • relu_encode_intEncoding of an integer as a Bool word: a sign bit followed by the binary encoding of its absolute value.
  • relu_encode_natBinary encoding of a natural number as a self-delimiting Bool word: each base-2 digit, least significant first, is preceded by a true…
  • op75_final_template_prefixA fixed Boolean word used as a fragment of the explicit network-encoding compiler.
  • ReLUHiddenLayerA hidden layer of width k consists of a k-by-k rational weight matrix and a rational bias vector of length k.
  • LinearOrder.toTotalPreorderEvery LinearOrder is a TotalPreorder.
  • TotalPreorderA total preorder: a preorder where ≤ is total (complete). This is weaker than LinearOrder — it does NOT require antisymmetry or…

Lean source view module on GitHub

theorem op75_final_encoding_template (t : ℕ) :
    true :: relu_encode_hidden_layer3 (op75_compact_final t) =
      op75_final_template_prefix ++ relu_encode_rat (-(t : ℚ)) ++ ([] : List Bool) ++
        relu_encode_rat (t : ℚ) ++ ([] : List Bool) := by
  simp [op75_compact_final, op75_compact_layer, relu_encode_hidden_layer3, relu_encode_vec3,
    relu_encode_rat, relu_encode_int, relu_encode_nat, op75_final_template_prefix, List.append_assoc] <;> norm_num <;> rfl
See it in the graphDiscuss this resultChallenge it