INDEX

Explanations

pure followed by noun

The neuron fires on occurrences of the word “pure” (in any casing).

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

is

-3.53

 they

-3.34

-3.22

ৢ

-2.97

↵↵

-2.88

to

-2.86

栳

-2.75

ers

-2.58

幸好

-2.56

 dritte

-2.53

POSITIVE LOGITS

Despite

3.50

While

3.47

After

3.20

Unlike

3.13

When

3.11

It

3.09

With

3.06

Although

3.00

Ironically

2.98

 北市

2.95

Activations Density 0.005%