INDEX

Explanations

overt, open, explicit

This neuron detects discourse markers and emphatic adverbs (and related punctuation) that signal contrast, emphasis, or framing 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

-1.57

</

-1.38

-1.29

providedIn

-1.27

ALSO

-1.25

 mucho

-1.24

 năm

-1.21

 statut

-1.20

 значит

-1.20

あまり

-1.19

POSITIVE LOGITS

 shown

1.37

 wüns

1.28

 stated

1.28

 involvement

1.27

but

1.27

you

1.27

几种

1.23

をしています

1.23

 mentioned

1.23

 přímo

1.22

Activations Density 0.052%