INDEX

Explanations

ex followed by common word endings

The neuron fires on any token containing the substring “ex.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ǿ

-1.35

äld

-1.26

 gladly

-1.25

 pedras

-1.25

 receber

-1.24

<td>

-1.23

-1.22

arschu

-1.20

瑞典

-1.18

 moedas

-1.17

POSITIVE LOGITS

wob

1.61

神色

1.49

 größer

1.42

kowania

1.42

浑身

1.41

EX

1.41

MONO

1.41

uarine

1.37

 leyendo

1.36

es

1.33

Activations Density 0.056%