INDEX

Explanations

planeta, relacionamento

This neuron detects non-English (especially Romance‐language) words.

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

 која

2.73

 которая

2.56

 która

2.52

 якая

2.50

 koja

2.48

 která

2.41

 която

2.36

 яка

2.31

 herself

2.23

 ktorá

2.23

POSITIVE LOGITS

 который

2.48

 який

2.44

 ktorý

2.36

 koji

2.34

 který

2.27

 који

2.25

 който

2.09

 झाली

2.06

 કરવી

2.03

 केली

1.93

Activations Density 0.151%