INDEX

Explanations

PlayStation Mobile

The neuron responds to unusually long, multi-syllable tokens (often technical or domain-specific terms).

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

 πρέπει

-1.31

 pertinentes

-1.30

 должен

-1.30

⫹

-1.29

*}$

-1.29

 vermel

-1.27

 sár

-1.26

𝑛

-1.26

*}\

-1.25

/}

-1.23

POSITIVE LOGITS

 other

1.74

 vittoria

1.46

 lainnya

1.45

 next

1.42

 which

1.42

its

1.40

 beginning

1.35

Noter

1.31

 другая

1.30

 svůj

1.30

Activations Density 0.231%