INDEX

Explanations

gradient

The main thing this neuron does is detect mentions of “gradient” (and related gradient‐function keywords like “linear‐gradient”).

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

-1.61

 needs

-1.53

 leggera

-1.50

主要是

-1.48

is

-1.45

 необходимости

-1.45

 damaligen

-1.43

Рецепт

-1.43

著名

-1.43

him

-1.43

POSITIVE LOGITS

of

1.77

ÑOS

1.74

 deſt

1.53

woordig

1.52

 réus

1.50

 ребята

1.49

bigcirc

1.47

ла

1.42

 تبد

1.41

 いた

1.41

Activations Density 0.010%