INDEX

Explanations

code contexts and keywords

The neuron activates on numeric tokens (literal numbers or numeric sequences) 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

 which

-1.16

 where

-1.10

the

-1.05

 rozwiąz

-0.97

iliki

-0.96

によっては

-0.94

yn

-0.94

 This

-0.93

malte

-0.93

 Этот

-0.91

POSITIVE LOGITS

 چیست

1.23

viate

1.23

.(*

1.21

ってみて

1.20

崭

1.19

やろ

1.13

ométrique

1.12

iramente

1.10

 Glou

1.09

來的

1.09

Activations Density 0.005%