INDEX

Explanations

things that are not

The neuron fires on occurrences of the word “not,” i.e. it detects negation.

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

of

-2.33

to

-2.16

for

-2.00

通过

-1.97

-1.88

at

-1.83

 after

-1.71

with

-1.70



-1.66

 melalui

-1.63

POSITIVE LOGITS

哢

2.06

 choisissez

1.91

 insole

1.89

 consultez

1.88

 akin

1.85

 located

1.84

 situated

1.83

他才

1.80

1.78

 большинстве

1.77

Activations Density 0.050%