INDEX

Explanations

secure hashes and versions

This neuron detects the numeric sequence “2-5-6” (e.g. the “256” in “SHA-256,” version numbers, or similar digit runs).

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

-2.05

 from

-1.92

it

-1.91

 because

-1.83

 even

-1.73

and

-1.73

 should

-1.72

 seems

-1.71

for

-1.69

 during

-1.69

POSITIVE LOGITS

 wondrous

1.48

 marvelous

1.46

 myriad

1.46

tych

1.43

 stalwart

1.42

 CURRENT

1.41

 magnific

1.40

 delightfully

1.39

 Dinas

1.38

柔ら

1.38

Activations Density 0.012%