INDEX

Explanations

mathematical expressions and greek letters

This neuron does not activate on any tokens—it doesn’t detect any pattern in the provided 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

 bemerken

-1.79

-1.70

 ofreció

-1.52

can

-1.38

之意

-1.37

大幅

-1.33

 sellado

-1.32

 acoged

-1.31

 obecnie

-1.31

-1.30

POSITIVE LOGITS

 familiari

1.39

Ⴈ

1.37

↵

1.36

يست

1.32

 знаю

1.29

 familiare

1.27

 этом

1.24

 delicioso

1.23

šin

1.21

%。

1.20

Activations Density 0.019%