INDEX

Explanations

The neuron is responding to very high‐frequency, common English words (function words and simple content words like in, has, its, is, just, matter, etc.).

"various" or "different"

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

vább

-1.32

傚

-1.25

 TOPOLOGY

-1.23

 Poincar

-1.23

rably

-1.23

וּ

-1.20

ượu

-1.19

 Olivenöl

-1.18

aterally

-1.17

 смогут

-1.17

POSITIVE LOGITS

1.26

—

1.24

 разных

1.20

贷款

1.19

 różnych

1.18

 verschied

1.17

 verschiedene

1.15

 разные

1.13

 beeindruck

1.09

 verschiedenen

1.07

Activations Density 0.028%