INDEX

Explanations

not much of anything

The neuron specifically fires on occurrences of the degree adverb “much.”

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

 somewhat

-1.30

 motoru

-1.22

 både

-1.20

 konfigura

-1.20

somewhat

-1.19

 již

-1.17

 pérd

-1.17

 zowel

-1.16

឵

-1.13

both

-1.13

POSITIVE LOGITS

of

1.72

 else

1.72

at

1.38

 After

1.34

if

1.30

 except

1.30

 difference

1.22

 Unlike

1.20

 Besides

1.16

Unlike

1.15

Activations Density 0.011%