INDEX

Explanations

text editor features

The neuron activates on code‐style constructs (e.g. keywords, directives, imports, punctuation) and thus identifies lines of source code or code snippets.

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

 сборной

-1.51

 then

-1.38

if

-1.34

 представители

-1.31

 such

-1.30

 detalló

-1.30

 telefonisch

-1.29

for

-1.27

 price

-1.25

 indicó

-1.23

POSITIVE LOGITS

ſa

1.51

 espirituales

1.41

LoadScene

1.40

 jScrollPane

1.34

vaných

1.28

ÂN

1.28

 σελ

1.23

IATION

1.23

 Theologie

1.22

 править

1.22

Activations Density 0.283%