INDEX

Explanations

explaining certainty or cause

The neuron fires on adverbial intensifiers that express certainty or obviousness (e.g. “clearly,” “apparently,” “obviously”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

-1.50

â

-1.45

–

-1.42

of

-1.41

-1.38

Ռ

-1.29

 طويلة

-1.28

了一些

-1.28

分辨

-1.21

 Soph

-1.20

POSITIVE LOGITS

 different

1.76

 another

1.73

 lepší

1.65

 znalaz

1.62

 Hinblick

1.55

camos

1.54

 todays

1.51

 забезпе

1.44

 wygod

1.44

 diferentes

1.43

Activations Density 0.003%