INDEX

Explanations

will be + completion/state

The neuron primarily activates on numeric tokens (i.e. standalone numbers) in the text.

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

 kysy

-2.13

炅

-1.98

おきます

-1.98

«,

-1.91

喈

-1.88

 vaih

-1.85

issait

-1.84

 houſe

-1.84

 mitään

-1.80

咴

-1.80

POSITIVE LOGITS

2.27

1.84

);

1.79

が開

1.77

Set

1.76

percent

1.71

筆記

1.69

Get

1.68

 Punkte

1.68

 Will

1.67

Activations Density 0.048%