INDEX

Explanations

names

The neuron strongly responds to capitalized tokens that are parts of people’s names.

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

StoryboardSegue

-2.03

"]));

-2.02

ーパー

-1.94

ೢ

-1.84

蛱

-1.77

lismo

-1.77

rightarrow

-1.76

AdvEx

-1.74

গ

-1.74

ariado

-1.73

POSITIVE LOGITS

They

2.03

 that

1.90

 this

1.82

if

1.63

 umożli

1.59

 which

1.58

—

1.57

 любого

1.47

The

1.42

 größeren

1.41

Activations Density 0.456%