INDEX

Explanations

numbers and dates

The neuron fires on numeric literal values (especially decimal and integer settings) in configuration or data files.

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

 also

-1.82

-1.71

-1.64

elif

-1.61

-1.50

 hebt

-1.40

 môžete

-1.38

Ihr

-1.34

 возника

-1.33

or

-1.32

POSITIVE LOGITS

こんな感じ

1.56

 vendar

1.55

ダイソー

1.52

PARATION

1.48

 idyllic

1.44

 leis

1.41

 anser

1.40

 revolu

1.38

 telah

1.34

無印

1.34

Activations Density 0.066%