INDEX

Explanations

average statistics

The neuron activates on numeric tokens, effectively detecting numbers and numerical measurements in the text.

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

ก็

-1.43

of

-1.37

on

-1.34

for

-1.21

 also

-1.21

水

-1.16

=”

-1.04

‘

-1.04

她

-1.03

to

-1.01

POSITIVE LOGITS

Zubereitung

1.40

凄く

1.27

ことがある

1.21

ਗ਼

1.20

 sorprender

1.14

Ceci

1.14

Woah

1.13

Jegy

1.12

絶妙

1.12

 TAMBÉM

1.10

Activations Density 0.089%