INDEX

Explanations

notes at the end

The neuron principally spots numeric tokens (e.g. integers or decimal numbers).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Despite

0.57

Despite

0.55

But

0.54

 Although

0.54

Although

0.54

why

0.54

なぜ

0.52

But

0.52

िसिटी

0.51

 which

0.51

POSITIVE LOGITS

[]');

0.59

|_

0.54

IRST

0.52

([(

0.51

ОВА

0.51

$.\

0.51

ined

0.50

 व्हेल

0.49

&_

0.49

[{\

0.49

Activations Density 0.028%