INDEX

Explanations

if statements

This neuron detects programming‐language syntax tokens—especially control‐flow and include keywords (e.g. “if,” “require/require_once”) and code delimiters or tags (e.g. “<?,” “<”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

<bos>

-2.14

-1.98

Etimología

-1.95

 parku

-1.89

 martie

-1.87

ጺ

-1.84

);*/

-1.82

dział

-1.80

تیجه

-1.79

 própri

-1.78

POSITIVE LOGITS

↵↵

2.42

it

2.23

又は

2.14

’,

1.86

 hoặc

1.85

1.80

is

1.80

 или

1.80

Activations Density 0.249%