INDEX

Explanations

explaining reasons and consequence

The neuron selectively detects high‐level auxiliary and copular verbs (and similar modals) such as “is,” “are,” “must,” “need,” and “called.”

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

 such

-1.53

 still

-1.30

 lembrar

-1.27

the

-1.25

ukraine

-1.15

 rosto

-1.10

};*/

-1.10

 ancora

-1.10

ProgressHUD

-1.05

jauh

-1.05

POSITIVE LOGITS

 naturligt

1.47

 istället

1.35

 също

1.29

 ögon

1.28

龇

1.26

而不是

1.21

 ingenting

1.21

 також

1.20

uellement

1.19

 fredag

1.19

Activations Density 0.153%