INDEX

Explanations

primary or related keywords followed by descriptive terms

The neuron flags occurrences of the words “primary” (and to a lesser extent “secondary”) when they’re used as class or label modifiers.

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

</strong>

-3.06

 fabulous

-2.06

-2.05

´

-2.03

‘

-1.98

 […]

-1.94

 magnificent

-1.94

…

-1.91

 tevens

-1.88

 superb

-1.87

POSITIVE LOGITS

If

2.09

滙

2.08

蛻

2.02

 शुर

1.99

𓋼

1.97

鏵

1.97

產業

1.92

”،

1.91

鈎

1.88

軼

1.87

Activations Density 0.000%