INDEX

Explanations

its followed by diverse words

The neuron specifically detects the word “its” (including the contraction “it’s”).

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

</td>

-1.62

on

-1.53

-1.52

もいい

-1.52

,”

-1.50

}$,

-1.40

or

-1.37

-1.36

âr

-1.35

 Eröffnung

-1.34

POSITIVE LOGITS

drei

1.67

 bedeckt

1.60

 jugado

1.59

鹀

1.57



1.52

 refirió

1.50

\*

1.48

鰨

1.47

 göm

1.45

biendo

1.45

Activations Density 0.024%