INDEX

Explanations

start of periods

The neuron is highly selective for the word “inception” (especially when used in date‐range contexts).

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

筇

-2.59

 сказала

-2.52

塤

-2.33

layed

-2.31

」「

-2.27

</h3>

-2.25

...";

-2.20

</th>

-2.20

鱚

-2.20

">>

-2.16

POSITIVE LOGITS

Now

2.73

潆

2.50

2.42

The

2.28

ing

2.22

沕

2.16

pecabe

2.09

lovakia

2.06

梠

2.05

 sería

2.03

Activations Density 0.003%