INDEX

Explanations

code component lifecycle methods

The neuron fires most strongly on tokens that are API method or event names (often CamelCase or dollar‐prefixed), e.g. “valueChanges,” “value_counts,” “$on,” “subscribe,” “composer,” 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

</h3>

-1.76

 großzü

-1.72

 mismos

-1.59

暃

-1.52

 Gebä

-1.52

 Geschä

-1.52

 Räume

-1.48

乄

-1.48

 مشابه

-1.48

ített

-1.47

POSITIVE LOGITS

1.53

 denunci

1.42

 politi

1.34

 responsabili

1.32

 glau

1.29

 Fichier

1.28

🗺

1.28

 sene

1.27

 temperat

1.27

 reis

1.27

Activations Density 0.004%