threshold-xor3-mag14

3-input XOR using cascade architecture with mag-7 XOR blocks.

Note: A flat 3-hidden architecture achieves magnitude 10. See threshold-xor3-mag10.

Architecture

a  b           c
│  │           │
└──┴──┐        │
      ▼        │
   ┌─────┐     │
   │XOR-7│     │
   └──┬──┘     │
      │        │
      └────┬───┘
           ▼
        ┌─────┐
        │XOR-7│
        └──┬──┘
           ▼
      XOR3(a,b,c)

Parameters

Neurons	6 (3 per XOR)
Layers	4
Parameters	18
Magnitude	14 (2 × 7)

Comparison

Version	Magnitude	Architecture
xor3 (base)	20	Cascade, original XOR
xor3-mag14	14	Cascade, mag-7 XOR
xor3-mag10	10	Flat 3-hidden

Usage

from safetensors.torch import load_file

w = load_file('model.safetensors')
# See model.py for implementation

License

MIT