INDEX

Explanations

updates and new information

This neuron fires on mentions of “update” (and its variants like updates, updated, up‐date) indicating blog or content update announcements.

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

mā

-1.28

裁判

-1.27

 podremos

-1.25

 only

-1.24

 hogyan

-1.20

 appelez

-1.19



-1.17

lievo

-1.16

скан

-1.08

olak

-1.07

POSITIVE LOGITS

實在

1.32

ラーメン

1.32

这是一个

1.31

 chapitre

1.27

疍

1.26

 uppdater

1.25

 полноцен

1.24

 polícia

1.24

 incidente

1.23

↵↵↵↵↵↵↵↵↵↵↵

1.22

Activations Density 0.040%