INDEX

Explanations

arrow and chevron icons

The neuron strongly activates on UI‐related arrow indicator tokens (e.g. “chevron” or “arrow”), i.e. words naming arrow/chevron icons or CSS classes.

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

we

-1.34

would

-1.30

我們

-1.24

也有

-1.23

也会

-1.20

↵

-1.20

 masih

-1.16

 untuk

-1.14

 bagi

-1.11

POSITIVE LOGITS

ěř

1.55

じゃあ

1.27

以上です

1.24

どういう

1.23

蹑

1.21

 sklep

1.20

 TAMBIÉN

1.20

をご覧ください

1.18

อื่น

1.17

を教えて

1.17

Activations Density 0.014%