INDEX

Explanations

list or table headers

The neuron strongly activates on the negation contraction “don’t.”

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

................

1.02

================

1.00

----------------

0.99

することができる

0.97

################

0.97

ることができる

0.96

ㅋㅋㅋㅋㅋㅋㅋㅋ

0.96

0.94

________________

0.93

0.91

POSITIVE LOGITS

 wellknown

1.54

↵

1.31

ナトリ

1.28

 അദ്ദേ

1.23

 socalled

1.19

 、,

1.15

 andRow

1.13

$%

1.12

 സെക്ര

1.05

 acknow

1.05

Activations Density 0.000%