INDEX

Explanations

The neuron fires strongly on tokens naming specific technical or professional domains (e.g. “electricity,” “intelligence,” “health,” “information,” etc.), so it’s detecting references to particular fields or industries.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ｃ

0.82

With

0.79

My

0.79

Of

0.78

이

0.75

Ａ

0.74

Ｐ

0.72

Ｍ

0.72

Ｌ

0.72

Ｓ

0.71

POSITIVE LOGITS

 indústria

1.20

industrie

1.16

 промышленности

1.16

業界

1.13

行业的

1.06

 vállalat

1.05

 industry

1.04

 financement

1.00

 sector

0.98

 oblast

0.98

Activations Density 1.601%