INDEX

Explanations

Gujarat Rojgar Samachar

This neuron fires most strongly on unusually long or uncommon tokens (e.g. technical terms, long words, or numeric strings).

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

or

-1.78

on

-1.59

and

-1.50

 This

-1.48

然

-1.34

$}

-1.32

少ない

-1.32

 like

-1.31

 after

-1.30

 before

-1.28

POSITIVE LOGITS

 erhi

1.44

 illustrazione

1.43

⮑

1.42

 ส์

1.37

 collezione

1.34

fillment

1.33

 secrétaire

1.30

 veder

1.30

 jepang

1.28

 beautifully

1.27

Activations Density 0.016%