INDEX

Explanations

The neuron activates on numeric tokens—especially measurements or size values—i.e. any token composed of digits.

size of

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

ısından

-1.54

 expresses

-1.33

ﻊ

-1.30

 respald

-1.30

 behaves

-1.29

 играет

-1.24

!#

-1.23

Ỳ

-1.22

-1.18

俣

-1.18

POSITIVE LOGITS

in

1.70

In

1.36

usually

1.34

 pien

1.30

摆

1.30

 Veränderungen

1.26

 לאחר

1.23

</strong>

1.22

 wahrscheinlich

1.21

 男人

1.21

Activations Density 0.036%