INDEX

Explanations

code delimiters and mathematical symbols

The neuron activates on mathematical operations and numeric-processing tokens (e.g. arithmetic/bitwise operators like +, *, >>, %, and math functions like sin, sqrt, log).

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

are

-1.38

 they

-1.22

ไม่

-1.20

you

-1.14

 ۱۴۰۰

-1.06

rifugal

-1.04

などが

-1.02

 ciclo

-1.01

ข่าว

-1.00

="

-1.00

POSITIVE LOGITS

Then

1.03

 muft

1.03

 then

1.02

and

1.02

bezpe

1.01

幸好

1.00

感は

0.99

colli

0.99

贻

0.98

 некото

0.96

Activations Density 0.013%