INDEX

Explanations

either, multiple, or choice

The neuron fires strongly on words that express quantity, extent, or comparison (e.g. “multiple,” “more,” “much,” “same,” etc.).

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

 have

-1.58

 where

-1.48

be

-1.43

its

-1.43

 such

-1.41

-1.38

as

-1.31

 even

-1.28

その

-1.27

how

-1.25

POSITIVE LOGITS

equipped

1.49

either

1.38

usually

1.34

laria

1.33

brought

1.33

probably

1.31

Either

1.30

――――

1.30

どっち

1.30

↓↓

1.29

Activations Density 0.510%