INDEX

Explanations

initiating an action

The neuron fires on verbs that indicate the start of an action (e.g., “began,” “started”).

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

-2.52

no

-2.06

 each

-1.88

-1.76

 this

-1.75

it

-1.75

 then

-1.59

 every

-1.54

 there

-1.53

-1.52

POSITIVE LOGITS

蹓

1.43

szt

1.40

Źródło

1.32

 появля

1.31

Алексей

1.28

佧

1.28

xvi

1.27

↓↓

1.27

Immediately

1.26

őzés

1.26

Activations Density 0.042%