INDEX

Explanations

and / to followed by verb

The main thing this neuron does is find numeric tokens (numbers, including those with decimal points).

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

of

-1.75

if

-1.70

 gdzie

-1.38

無く

-1.32

(%)

-1.27

 hicieron

-1.26

 dimana

-1.24

");

-1.23

無い

-1.23

</h2>

-1.22

POSITIVE LOGITS

 mentales

1.49

совет

1.45

 “[

1.45

 Instead

1.45

1.43

irsten

1.40

 vœux

1.36

xelles

1.34

struzioni

1.34

 Presumably

1.34

Activations Density 0.072%