INDEX

Explanations

numeric identifiers and delimiters

The neuron fires on numeric tokens (digits) used as indices or identifiers in code.

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

ところに

-1.30

-1.23

 where

-1.22

-1.19

-1.18

-1.15

-1.13

ことができる

-1.12

になりたい

-1.12

POSITIVE LOGITS

】：

1.24

kym

1.23

<bos>

1.22

は

1.19

 restoran

1.19

 seksual

1.18

پول

1.17

 brillantes

1.17

勇者

1.16

และ

1.13

Activations Density 0.041%