INDEX

Explanations

background images and filenames

The neuron selectively activates on tokens containing sequences of digits (numeric string fragments).

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

 with

-1.63

for

-1.57

ּוֹ

-1.41

if

-1.34

ַּ

-1.28

at

-1.27

any

-1.19

 that

-1.19

its

-1.16

in

-1.13

POSITIVE LOGITS

fokus

1.34

 heureuse

1.28

identifikasi

1.23

bilden

1.23

 Институ

1.20

ﻻ

1.20

ⅰ

1.18

ׇ

1.17

Deadpool

1.15

⽼

1.15

Activations Density 0.005%