INDEX

Explanations

code parsing and execution

The neuron fires on calls to load/parse functions that deserialize structured data (e.g. YAML.load, JSON.loads/Parse, File.read→load, etc.).

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

-1.37

<strong>

-1.36

ppings

-1.31

着一个

-1.30

not

-1.29

™

-1.27

 oraz

-1.23

了个

-1.21

 Spectrometry

-1.21

↵↵

-1.20

POSITIVE LOGITS

 that

1.63

imbang

1.50

jauh

1.44

cepatan

1.38

お問い合わせ

1.34

 from

1.34

新型コロナ

1.31

of

1.29

スズキ

1.27

顯示

1.27

Activations Density 0.023%