INDEX

Explanations

route to

The main thing this neuron does is pick out second-person imperative verbs giving step-by-step instructions (e.g., “head,” “jump,” “enter”).

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

 that

-2.66

が違う

-1.73

ğin

-1.63

 those

-1.57

 pisau

-1.51

 иногда

-1.49

 topik

-1.49

 which

-1.45

んだから

-1.44

gridColumn

-1.43

POSITIVE LOGITS

áles

1.45

謚

1.43

 τὴν

1.32

 your

1.30

 жидкости

1.27

leste

1.27

 ティー

1.27

他在

1.26

<blockquote>

1.24

...

1.24

Activations Density 0.058%