INDEX

Explanations

questions and answers

The neuron fires on large runs of whitespace or indentation characters (i.e. long sequences of space/tab tokens).

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

-2.31

-2.03

-1.99

-1.77

各样的

-1.69

 cuyos

-1.55

しました

-1.54

就算是

-1.50

 przede

-1.48

 idol

-1.45

POSITIVE LOGITS

ременно

1.78

ignment

1.76

sthe

1.73

 内衣

1.73

这

1.68

 biodivers

1.64

ματα

1.63

ization

1.63

 transporta

1.61

 新款

1.59

Activations Density 0.028%