INDEX

Explanations

thousands and millions

The neuron activates on numeric expressions—digits, number words, and quantitative tokens.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

for

-1.89

 during

-1.59

ในการ

-1.58

 when

-1.55

tefan

-1.42

根据

-1.38

于是

-1.33

大家都

-1.30

two

-1.28

وجه

-1.28

POSITIVE LOGITS

曏

1.32

izarse

1.27

 asimismo

1.26

くれます

1.23

躥

1.20

</strong>

1.19

 varier

1.16

 角色

1.16

Also

1.15

 consulter

1.14

Activations Density 0.096%