INDEX

Explanations

or other options

The neuron activates on tokens that include or represent numeric counts (i.e. words with appended digits).

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

 بكل

-0.94

只是一个

-0.89

 КУ

-0.88

zusetzen

-0.87

esthes

-0.87

Abigail

-0.85

 keinerlei

-0.84

 vermag

-0.84

 resumen

-0.84

Besch

-0.84

POSITIVE LOGITS

or

1.09

における

0.94

での

0.93

 alebo

0.92

 atau

0.91

 because

0.85

][/

0.84

 with

0.84

\\

0.82



0.82

Activations Density 0.276%