INDEX

Explanations

[Cite

The neuron fires on the start of a LaTeX superscript (the “^{” that begins an exponent in math mode).

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

 from

-2.50

in

-2.39

But

-2.20

だけど

-2.11

方針

-2.05

着的

-2.02

he

-1.96

などは

-1.95

柔和

-1.89

by

-1.85

POSITIVE LOGITS

2.56

 vile

2.50

 allmäh

2.41

 erheblich

2.41

那你

2.39

 häufiger

2.33

 vorüber

2.31

 nachfol

2.28

ञ

2.16

 โดย

2.14

Activations Density 0.003%