INDEX

Explanations

mean squared error, root mean square

The neuron activates on numeric expressions (digits, decimal numbers, and related measurement terms) in the text.

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

who

-1.77

 nhưng

-1.68

 αλλά

-1.66

 fearful

-1.59

嗰

-1.54

 Kafka

-1.51

 valentín

-1.48

鶩

-1.48

 Spearman

-1.45

并没有

-1.41

POSITIVE LOGITS

1.63

1.38

if

1.37

ing

1.37

 traditionelle

1.36

 podríamos

1.33

刽

1.30

due

1.27

 depending

1.22

Activations Density 0.005%