INDEX

Explanations

how much questions

The neuron flags price-query phrases—i.e. questions of the form “how much/does/cost [X]” or other explicit cost/amount inquiries.

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

bildet

-1.10

 traseira

-1.07

 toute

-1.06

انيف

-1.03

ഞ

-1.02

bluza

-1.02

鉨

-1.00

そんな

-0.98

ketten

-0.95

Lorem

-0.94

POSITIVE LOGITS

 each

1.89

 poszczegól

1.58

 average

1.27

娅

1.24

in

1.20

 verschiedenen

1.19

 various

1.19

 these

1.17

 verschillende

1.16

 sebenarnya

1.11

Activations Density 0.029%