INDEX

Explanations

repetition for intensification

The neuron fires on comparative degree words—adjectives and adverbs like “smaller,” “more…,” or “distant” that express relative change.

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

-2.05

-1.68

-1.63

不仅

-1.62

is

-1.50

 muchos

-1.46

すべての

-1.44

 trusty

-1.44

 conhe

-1.43

 MUCH

-1.39

POSITIVE LOGITS

➺

1.50

lijden

1.50

わけです

1.44

 DEPOSIT

1.43

utentic

1.42

 kroppen

1.41

鮡

1.41

سجيلات

1.39

because

1.34

と感じ

1.34

Activations Density 0.005%