INDEX

Explanations

data labels, errors, or identifiers

The neuron fires on vertical-bar characters (“|”), i.e. column or field separators in tables and code listings.

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

 แล้ว

-1.38

</em>

-1.34

Cri

-1.32

 культура

-1.27

kör

-1.25

駐車

-1.23

-1.17

駆け

-1.17

 fiscales

-1.17

POSITIVE LOGITS

 fakty

1.87

咴

1.64

・・・」

1.59

 kupić

1.59

 debole

1.57

༑

1.56

乄

1.53

 bukanlah

1.50

ellees

1.48

 difficoltà

1.45

Activations Density 0.082%