INDEX

Explanations

flick and flicker

The main thing this neuron does is detect occurrences of the substring “flick” in words.

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

на

-3.06

焢

-2.91

ve

-2.86

In

-2.78

ꦫ

-2.77

 exacting

-2.61

 analogous

-2.36

糶

-2.33

actéristi

-2.23

秝

-2.23

POSITIVE LOGITS

5.06

2.94

2.92

anyeol

2.77

the

2.42

!"

2.42

然后

2.42

地的

2.39

 plaatsen

2.30

gefügt

2.27

Activations Density 0.001%