INDEX

Explanations

I contractions

The neuron consistently fires on the very first content token of each text block (e.g. the opening word or header of a post or paragraph).

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

 then

-1.28

maksi

-1.27

 thereafter

-1.26

and

-1.26

וקה

-1.24

 thence

-1.23

ِّ

-1.23

 dieser

-1.21

 sula

-1.20

🪛

-1.19

POSITIVE LOGITS

It

1.85

you

1.56

ݯ

1.44

 musisz

1.41

 ponieważ

1.38

 случаев

1.35

 Mejora

1.35

 маленьких

1.33

 permanecer

1.29

 vence

1.27

Activations Density 0.216%