INDEX

Explanations

unicode_literals in Python code

The neuron fires on code tokens that specify or demand literal values or assignments—e.g. boolean/number literals (`true`, `0`), the assignment operator (`=`), and PropTypes “isRequired” flags.

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

-1.51

to

-1.45

-1.41

if

-1.36

-1.35

 endwhile

-1.34

</tfoot>

-1.33

 OGSÅ

-1.30

 this

-1.27

our

-1.25

POSITIVE LOGITS

 nettoyer

1.66

↵

1.55

<bos>

1.53

 koel

1.41

 empêcher

1.40

iny

1.40

 biedt

1.37

 arrêter

1.35

 goû

1.33

能在

1.33

Activations Density 0.001%