INDEX

Explanations

widely accepted as

The main thing this neuron does is detect phrases asserting broad consensus or recognition (e.g. “widely accepted,” “generally accepted,” “widely regarded as,” “widely considered as”).

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

᯾

-1.71

もちろん

-1.59

୩

-1.55

 поскольку

-1.54

 each

-1.50

 considerados

-1.48

you

-1.48

 ciencias

-1.45

 jika

-1.44

 introduce

-1.41

POSITIVE LOGITS

勠

1.92

蕯

1.80

1.65

鹮

1.56

 dekoratif

1.51

stroom

1.40

 اما

1.39

»:

1.38

就是

1.37

）

1.35

Activations Density 0.048%