INDEX

Explanations

demographic groups

The neuron detects numerical statistics (counts, percentages, and sample‐size figures) in the text.

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

 beendet

-1.35

ほら

-1.27

ようで

-1.24

ことも

-1.23

 raccol

-1.16

 attore

-1.16

にします

-1.13

嵓

-1.13

 rivel

-1.12

#+#

-1.11

POSITIVE LOGITS

 have

1.79

are

1.67

in

1.64

 were

1.41

がありますが

1.32

が

1.26

 will

1.10

は

1.09

薔薇

1.09

is

1.08

Activations Density 0.049%