INDEX

Explanations

Japanese particles and sentence endings

This neuron detects passages written in Japanese (non-Latin characters).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

ܗ

1.45

 prophets

1.43

 Applying

1.42

 metaphysics

1.42

 ବି

1.39

ylsulfanyl

1.35

adə

1.35

 diminishes

1.34

 polynomials

1.33

eningkatan

1.32

POSITIVE LOGITS

な

1.94

リ

1.94

バ

1.93

「

1.90

さ

1.86

じ

1.85

ガ

1.85

この

1.81

セ

1.81

コン

1.79

Activations Density 0.010%