INDEX

Explanations

specific phrases and topics

The neuron detects floating‐point numbers (numeric tokens containing decimal points).

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

);

-2.11

out

-1.67

];

-1.54

 countless

-1.35

its

-1.31

-1.30

 many

-1.30

 οποίο

-1.28

 intricate

-1.27

</h4>

-1.24

POSITIVE LOGITS

Зно

1.45

 bombar

1.44

 opini

1.40

 dore

1.37

 Indon

1.34

 nazis

1.34

talen

1.33



1.31

 scrivere

1.30

 latine

1.29

Activations Density 0.180%