File size: 1,192 Bytes
ccee2d3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
---
license: mit
tags:
- pytorch
- safetensors
- threshold-logic
- neuromorphic
---

# threshold-signbit4

Extract sign bit (MSB) of 4-bit number.

## Function

signbit4(a3, a2, a1, a0) = a3

In 2's complement representation:
- Output 0: non-negative (0 to 7)
- Output 1: negative (-8 to -1)

## Truth Table

| a3 | a2 | a1 | a0 | unsigned | signed | signbit |
|----|----|----|----|---------:|-------:|:-------:|
| 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 0 | 1 | 1 | 1 | 7 | +7 | 0 |
| 1 | 0 | 0 | 0 | 8 | -8 | 1 |
| 1 | 1 | 1 | 1 | 15 | -1 | 1 |

## Architecture

Single neuron extracting MSB:

- Weights: [1, 0, 0, 0]
- Bias: -1

Fires when: a3 - 1 >= 0, i.e., a3 >= 1

## Parameters

| | |
|---|---|
| Inputs | 4 |
| Outputs | 1 |
| Neurons | 1 |
| Layers | 1 |
| Parameters | 5 |
| Magnitude | 2 |

## Usage

```python
from safetensors.torch import load_file
import torch

w = load_file('model.safetensors')

def signbit4(a3, a2, a1, a0):
    inp = torch.tensor([float(a3), float(a2), float(a1), float(a0)])
    return int((inp @ w['neuron.weight'].T + w['neuron.bias'] >= 0).item())

print(signbit4(0, 1, 1, 1))  # 0 (value = +7)
print(signbit4(1, 0, 0, 0))  # 1 (value = -8)
```

## License

MIT