INDEX

Explanations

foreign language beginnings

The neuron fires on tokens containing non-ASCII characters or diacritical marks (i.e. accented or special‐letter 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

-2.03

螓

-1.91

 baden

-1.90

 cujo

-1.90

iss

-1.89

 reprodução

-1.88

<bos>

-1.88

 francus

-1.86

 cuja

-1.86



-1.85

POSITIVE LOGITS

↵

2.48

there

2.39

and

2.39

縝

2.30

or

2.22

 こう

2.02

轶

1.98

))$_

1.93

ГЛА

1.92

熒

1.88

Activations Density 0.152%