INDEX

Explanations

courtesy of

This neuron detects image‐credit attributions, especially the “Courtesy of” photo credit phrases.

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

 famosas

-2.31

LayoutConstraint

-2.22

⎚

-2.20

 Voraussetzung

-2.14

thschild

-2.03

üsseldorf

-2.02

 únicas

-1.98

 conocidas

-1.96

jenner

-1.92

嘥

-1.92

POSITIVE LOGITS

of

3.47

to

2.69

芣

2.59

2.56

This

2.50

2.47

2.34

྅

2.28

 lelaki

2.19

她是

2.16

Activations Density 0.003%