INDEX

Explanations

siege, blockade, cordon

The neuron fires on words describing siege‐like military or protest actions (e.g. blockade, besiege, surround, storm, block).

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

人們

-1.67

 wußte

-1.49

 erzä

-1.48

 erlä

-1.45

 あと

-1.43

ologous

-1.38

 Fácil

-1.38

一張

-1.36

 Zuhause

-1.34

POSITIVE LOGITS

 één

1.47

1.41

 único

1.37

 ihrer

1.36

of

1.34

 faisant

1.30

 двумя

1.29

 republiky

1.28

tics

1.27

 opnieuw

1.27

Activations Density 0.022%