INDEX

Explanations

titles, chapter headings, or questions

The neuron strongly responds to the literal empty‐string token (two single quotes: ‘’).

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

 начин

-1.50

stoffen

-1.42

 maquilla

-1.30

多吃

-1.24

 môže

-1.23

 มี

-1.23

moderne

-1.23

ITEN

-1.22

//!

-1.16

 türkiye

-1.16

POSITIVE LOGITS

ご覧

1.22

1.21

𝒾

1.15

 blijkt

1.13

hline

1.13

 which

1.13

but

1.11

 this

1.06

1.02

 Clim

1.02

Activations Density 0.000%