INDEX

Explanations

links and emails

The neuron specializes in spotting hyperlink and markup syntax (e.g. HTML tags and Markdown link brackets/parentheses).

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

RESUMEN

-1.70

 Eigenschaften

-1.59

 داشتم

-1.48

是一款

-1.41

trzyma

-1.39

きましょう

-1.38

SUND

-1.37

 نداره

-1.37

大幅

-1.34

dây

-1.34

POSITIVE LOGITS

the

2.30

our

1.79

to

1.73

or

1.51

:][

1.44

out

1.43

1.35

してみてください

1.34

or

1.31

 there

1.27

Activations Density 0.016%