INDEX

Explanations

Head, Spotlight, Br, var

The neuron strongly activates on standalone Arabic numerals (page numbers, years, counts, etc.).

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

in

-1.73

on

-1.49

 that

-1.48

 such

-1.38

to

-1.36

 with

-1.30

 other

-1.23

at

-1.17

any

-1.16

 primera

-1.16

POSITIVE LOGITS

1.54

“

1.31

TECN

1.27

这是一

1.20

了许多

1.13

 zatím

1.12

SEGU

1.10

 hôtels

1.10

pulseira

1.10

Información

1.09

Activations Density 0.001%