INDEX

Explanations

ending in uz

The neuron flags any token containing the letter sequence “uz.”

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

-2.66

.”

-2.61

锭

-2.31

是最

-2.25

 ujedno

-2.20

～！

-2.19

),

-2.19

 That

-2.16

》：

-2.16

絰

-2.13

POSITIVE LOGITS

3.05

亹

2.78

嘬

2.59

 they

2.52

ain

2.42

 évaluation

2.42

have

2.41

<blockquote>

2.38

ܙ

2.38

geſ

2.36

Activations Density 0.003%