INDEX

Explanations

programming, lists, names, foreign words

The neuron is picking out punctuation and delimiter tokens—things like parentheses, commas, back‐ticks, quotes, etc.

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

-1.67

 their

-1.34

}{\

-1.25

ก็

-1.18

 摄影

-1.17

its

-1.16

or

-1.15

and

-1.13

 With

-1.13

get

-1.12

POSITIVE LOGITS

 вами

1.36

ностях

1.27

 дополнитель

1.26

istaa

1.22

 erwähnten

1.21

adomo

1.20

 můžete

1.20

PENUTUP

1.16

 вашего

1.16

 แต่

1.15

Activations Density 0.174%