INDEX

Explanations

companies and organizations

The neuron activates on names of companies or other organizations.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 chưa

-0.99

 zatím

-0.98

„

-0.96

-0.93

んだよ

-0.87

 überprüfen

-0.86

</em>

-0.85

-0.84

POSITIVE LOGITS

 before

1.77

 where

1.47

before

1.42

 cuando

1.40

 sebelum

1.37

 when

1.34

and

1.25

 years

1.21

 dulu

1.20

 BEFORE

1.18

Activations Density 0.053%