INDEX

Explanations

until

The neuron strongly fires on sentence‐internal discourse markers or transitional cue words (e.g. “until,” “then,” “just,” “seems,” “use,” “enjoy”) that structure or shift the flow of 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

 when

-1.85

vedad

-1.30

 cuando

-1.23

cuando

-1.20

 ketika

-1.20

يمه

-1.16

 quando

-1.15

なんですね

-1.15

 gän

-1.13

 When

-1.11

POSITIVE LOGITS

pe

1.31

”

1.29

+"&

1.21

 остальные

1.20

val

1.19

.”

1.16

 yayınları

1.16

 musikk

1.16

habited

1.16

-”

1.15

Activations Density 0.031%