INDEX

Explanations

specific items, events, or places

The neuron is triggered by numeric expressions (counts, measurements, dates, quantities) in the text.

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

or

-1.27

 więc

-1.25

 more

-1.20

之前

-1.10

 powierzch

-1.09

and

-1.05

 additional

-1.04

-1.03

-1.01

英国

-1.01

POSITIVE LOGITS

 that

1.63

を

1.22

for

1.16

bisous

1.16

émie

1.14

 Masalah

1.10

 from

1.09

ENED

1.08

もしか

1.05

 ujar

1.05

Activations Density 0.179%