INDEX

Explanations

method calls and code structure

This neuron detects code identifiers and punctuation in fully qualified names (e.g. class or method names and their dot separators).

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

lético

-1.59

ﮩ

-1.57

𝓗

-1.53

鉺

-1.53

Hapana

-1.47

 અને

-1.46

Hélène

-1.46

ingabe

-1.45

hlte

-1.43

 comunicaciones

-1.40

POSITIVE LOGITS

of

1.98

 used

1.70

 other

1.66

 first

1.61

use

1.60

1.58

1.56

get

1.52

we

1.51

<bos>

1.46

Activations Density 0.022%