INDEX

Explanations

living people

The neuron selectively fires on standalone numeric tokens (i.e. sequences of digits) and its activation equals the number represented by that token.

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

方針

-0.85

月から

-0.84

 enhanced

-0.77

 weirdly

-0.76

håll

-0.75

 strangely

-0.75

 clergymen

-0.73

・・。

-0.72

 oiled

-0.72

ごろ

-0.71

POSITIVE LOGITS

 Instead

0.99

 magnificent

0.89

instead

0.89

Instead

0.89

 тыс

0.86

mafia

0.86

 instead

0.85

nosa

0.84

 นิ้ว

0.84

intptr

0.82

Activations Density 0.001%