INDEX

Explanations

cannot show or use

The neuron fires on numeric expressions (especially multi‐digit or decimal numbers).

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

 όμως

-1.44

後記

-1.38

 where

-1.32

timberland

-1.28

 besonderen

-1.27

Види

-1.27

時に

-1.27

⚋

-1.22

collier

-1.21

tamaris

-1.20

POSITIVE LOGITS

窄

1.33

」

1.30

get

1.30

…»

1.29

٥

1.29

 første

1.24

kette

1.21

style

1.21

langs

1.21

:**

1.20

Activations Density 0.092%