INDEX

Explanations

opening quotation marks

The neuron is looking for colon characters (":" punctuation) in the text.

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

 including

-1.32

が

-1.25

当時は

-1.25

ちなみに

-1.24

<bos>

-1.23

 with

-1.23

そのため

-1.22

had

-1.21

 came

-1.20

 took

-1.18

POSITIVE LOGITS

“

1.80

1.73

‘

1.48

ldquo

1.41

 ſever

1.34

 participer

1.30

『

1.18

бол

1.17

 BETWEEN

1.17

worked

1.17

Activations Density 0.014%