INDEX

Explanations

separate reports, Liberal government, agents advertising

The neuron responds to content-carrying, domain-specific nouns and technical terms.

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

дите

-1.28

 with

-1.25

 "}";

-1.24

 jsme

-1.21

ceptable

-1.21

ggars

-1.20

 jaket

-1.20

)->

-1.17

lossen

-1.14

 eletr

-1.14

POSITIVE LOGITS

 other

1.55

and

1.46

其他

1.38

»,

1.34

 anderen

1.33

 другими

1.33

you

1.32

อื่น

1.30

 entweder

1.30

You

1.22

Activations Density 0.168%