INDEX

Explanations

This neuron detects command-line instructions and related shell commands (e.g. sudo, apt-get, ssh, gnome-terminal, su, etc.).

shell commands

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

our

-1.61

 incluso

-1.35

 даже

-1.34

เรา

-1.31

we

-1.24

 baseado

-1.19

 ofrecemos

-1.18

 мы

-1.18

 với

-1.17

 até

-1.16

POSITIVE LOGITS

 then

1.55

THISDAY

1.20

こいつ

1.09

توضیحات

1.08

素晴

1.05

＞＞

1.05

お前

1.04

Algemeen

1.03

 می‌گو

1.03

обще

1.03

Activations Density 0.311%