INDEX

Explanations

Cassandra or boolean values

The neuron spotlights numeric tokens—any standalone sequences of 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

â

-1.40

שְׁ

-1.27

 недостатки

-1.17

 perdón

-1.16

 häufigsten

-1.14

спользова

-1.14

美国的

-1.14

 puțin

-1.11

加拿

-1.10

-1.09

POSITIVE LOGITS

ਫ਼

1.83

立

1.73

ਖ਼

1.72

ਗ਼

1.66

來

1.58

數

1.55

ড়

1.50

ţei

1.50

度

1.42

利

1.34

Activations Density 45.004%