INDEX

Explanations

B followed by number or initial

The neuron activates on standalone numeric tokens—integers, decimals, hex codes, citation numbers, etc.

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

Архівовано

-1.30

えた

-1.23

見えない

-1.20

 técnicos

-1.20

 buque

-1.18

 asegur

-1.17

 hjælpe

-1.17

 ciebie

-1.16

 sikre

-1.16

ったら

-1.16

POSITIVE LOGITS

 ness

1.38

𖧵

1.36

’，

1.33

 anni

1.22

 expri

1.21

 यदि

1.20

 MORE

1.20

 experimentation

1.20

۱۹

1.18

,',

1.18

Activations Density 0.040%