INDEX

Explanations

Expert opinion:

The neuron activates on numeric tokens (digits or numbers), picking up counts, measurements, years, and other numeric values.

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

 Première

-1.13

 mikrofon

-1.08

èvent

-1.05

 могат

-1.04

 kilomètres

-1.02

poptosis

-1.00

まず

-1.00

 gjorde

-0.99

①

-0.98

Hozzá

-0.96

POSITIVE LOGITS

<eos>

1.98

it

1.30

↵↵↵↵↵↵↵↵↵↵

1.07

if

1.05

↵↵↵↵↵↵↵↵↵

1.00

 there

0.99

↵↵↵↵↵↵↵↵↵↵↵↵

0.99

↵↵↵↵↵↵↵↵↵↵↵↵↵

0.98

↵↵↵↵↵↵↵↵↵↵↵

0.98

www

0.98

Activations Density 0.073%