INDEX

Explanations

foreign language word parts

The neuron strongly activates on subword pieces of proper names—i.e. fragments of capitalized entity or product names.

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

of

-1.55

也非常

-1.27

We

-1.24

 know

-1.21

And

-1.20

Pozri

-1.18

閡

-1.18

おわりに

-1.18

見学

-1.18

get

-1.17

POSITIVE LOGITS

 ANCHE

1.37

他們

1.23

micos

1.20

 Kräuter

1.19

'$

1.19

 (!__

1.17

他的

1.16

 universitaire

1.16

to

1.14

 FISCAL

1.14

Activations Density 0.008%