INDEX

Explanations

initiate actions or commands

The neuron fires most strongly on words that signal a shift or transition—especially verbs and nouns denoting change, movement, or time (e.g. “change,” “time,” “back”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

लेकिन

0.73

but

0.70

albeit

0.66

 nelle

0.65

oodles

0.64

 mutta

0.64

最初は

0.64

мыми

0.63

 sambil

0.63

после

0.63

POSITIVE LOGITS

 чтоб

1.05

 чтобы

0.98

 để

0.97

to

0.95

 żeby

0.89

 whereby

0.85

 щоб

0.78

เพื่อ

0.76

 ताकि

0.76

 upang

0.75

Activations Density 0.312%