INDEX

Explanations

won or scored

The neuron is looking for numeric tokens (statistics, scores, and other number 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

 that

-1.38

including

-1.36

例えば

-1.23

 різних

-1.19

それは

-1.17

 which

-1.13

它的

-1.12

 навіть

-1.09

őtt

-1.07

these

-1.05

POSITIVE LOGITS

our

1.06

 venter

1.05

for

1.04

Biografi

1.04

 fanns

1.03

ступи

1.02

 เข

1.01

樞

1.01

вому

1.01

 роза

0.99

Activations Density 0.028%