INDEX

Explanations

necessary and sufficient conditions

This neuron specifically activates on occurrences of the word “necessary” (in any casing or context).

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

呶

-1.58

ası

-1.29

螃

-1.23

 sepa

-1.22

şk

-1.21

érieurs

-1.21

 principaux

-1.20

雱

-1.18



-1.18

izielle

-1.12

POSITIVE LOGITS

to

2.39

for

2.09

and

1.41

 для

1.25

but

1.20

 precaution

1.18

 necessary

1.18

↵

1.15

以便

1.14

に応じて

1.12

Activations Density 0.013%