INDEX

Explanations

logarithms and calculus notation

This neuron detects LaTeX math‐mode syntax tokens (backslashes and common math operators like “d” or “ln”).

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

 deviennent

-1.78

 různ

-1.73

 recebeu

-1.69

&-

-1.64

 bénéficient

-1.63

茑

-1.63

因为它

-1.63

 ทำ

-1.63

‚¬

-1.62

させて頂きます

-1.59

POSITIVE LOGITS

of

2.78

the

2.22

 their

2.17

 these

2.13

for

2.06

 from

2.06

all

2.05

 each

2.02

our

1.98

 them

1.94

Activations Density 0.003%