INDEX

Explanations

introducing new points

This neuron detects discourse‐level transition words or adverbs (e.g. “Furthermore,” “Also,” “Secondly,” “Therefore”) that introduce additional points or questions.

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

そこは

-1.25

 encantó

-1.13

んだから

-1.12

そりゃ

-1.08

}}}{\

-1.07

 szá

-1.00

idends

-0.99

lho

-0.98

 BECAUSE

-0.97

ᐳ

-0.95

POSITIVE LOGITS

 also

1.99

if

1.58

how

1.54

How

1.52

 também

1.46

 également

1.43

 question

1.36

 również

1.34

Also

1.31

 también

1.31

Activations Density 0.105%