INDEX

Explanations

code structure delimiters

This neuron responds to closing delimiters—e.g. the “>” of an HTML tag or a “}” ending a code block.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

to

-1.70

 Also

-1.15

as

-1.13

 also

-1.08

 that

-1.02

for

-1.01

了一会儿

-1.00

 being

-1.00

ederen

-1.00

 with

-0.96

POSITIVE LOGITS

 beide

1.51

Both

1.48

 begge

1.34

 ambos

1.34

 ambas

1.20

这两

1.20

 båda

1.20

Kesimpulan

1.16

 оба

1.16

这两个

1.14

Activations Density 0.012%