INDEX

Explanations

describing characteristics of subjects

This neuron is specialized to detect the word “characteristics” (often in titles or section headings).

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

."

-1.95

derdag

-1.77

鬈

-1.60

嫫

-1.59

 Informationen

-1.55

</i>

-1.54

In

-1.49

艽

-1.45

*}

-1.44

⎔

-1.44

POSITIVE LOGITS

 tämän

1.99

and

1.80

 tämä

1.75

 suuri

1.73

 tästä

1.62

 laajenta

1.60

 henkilö

1.58

 podem

1.55

 kaksi

1.53

にかく

1.52

Activations Density 0.139%