INDEX

Explanations

Don followed by names

The neuron fires on the proper name “Don.”

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

section

-2.98

-2.86

up

-2.67

of

-2.58

-2.56

re

-2.48

 عندما

-2.48

ၟ

-2.47

going

-2.45

 antics

-2.41

POSITIVE LOGITS

᯾

3.31

’.

3.25

 westfalen

2.98

 replete

2.95

’.”

2.84

’;

2.83

怍

2.78

悒

2.72

嚆

2.72

不过是

2.67

Activations Density 0.012%