INDEX

Explanations

list ellipsis [...]

The neuron fires strongly on numerical tokens or sequences of digits.

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

т

1.47

на

1.46

स

1.38

י

1.38

ди

1.38

1.30

न

1.21

1.17

क

1.16

в

1.13

POSITIVE LOGITS

ように

1.22

ंटिक

1.19

 parâmetros

1.17

 problemlos

1.15

 stifle

1.12

อยู่ที่

1.12

 terang

1.09

 antennis

1.05

 höchste

1.05

 Nessa

1.05

Activations Density 0.076%