INDEX

Explanations

The neuron strongly activates on import statements (lines beginning with “import”) in code files.

structural influence

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

 також

-1.23

》,

-1.15

 arrivant

-1.10

🛕

-1.07

también

-1.07

 vété

-1.06

侈

-1.06

horabuena

-1.05

"";

-1.04

izacion

-1.02

POSITIVE LOGITS

to

1.36

if

1.15

my

1.12

 into

1.09

そのまま

1.02

 можно

1.00

it

0.99

use

0.93

you

0.91

as

0.90

Activations Density 0.016%