INDEX

Explanations

companies and their services

The neuron activates on capitalized named entities—especially brand and product names (e.g., “Xbox,” “Apple,” “Facebook,” “Paytm,” 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

每

-1.32

𝐞

-1.31

 that

-1.24

ʀ

-1.23

 herramientas

-1.23

 adecuado

-1.20

It

-1.20

 nuevas

-1.20

üre

-1.19

С

-1.16

POSITIVE LOGITS

 while

1.54

if

1.39

 diferite

1.34

by

1.33

 великолеп

1.29

tisp

1.28

tahankan

1.27

 either

1.26

or

1.24

ragal

1.23

Activations Density 0.170%