INDEX

Explanations

editor of publications

The neuron detects mentions of editorial roles (e.g. “editor,” “Editor‐in‐Chief,” “edited”).

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

vantaged

-1.95

 überprü

-1.86

to

-1.78

 связанные

-1.75

 protože

-1.72

 utilizan

-1.71

其他的

-1.71

と言われる

-1.66

vinte

-1.66

 takže

-1.66

POSITIVE LOGITS

狆

2.33

 engels

2.16

猡

2.08

↵

2.03

ୈ

1.86

城的

1.84

 editors

1.79

 gruppen

1.78

 straff

1.78

 serien

1.77

Activations Density 0.020%