INDEX

Explanations

forms of "to be" verb

The neuron flags relatively informative content words (topical nouns, pronouns, and other mid‐frequency “keyword” tokens) rather than common function 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

does

0.79

Does

0.78

 включает

0.78

 does

0.77

 கொண்டுள்ளது

0.72

 DOES

0.71

 imao

0.69

 इसने

0.68

 avevano

0.65

 ചെയ്തി

0.64

POSITIVE LOGITS

 είναι

1.95

är

1.77

 adalah

1.59

are

1.54

yra

1.54

 jsou

1.52

 são

1.51

是

1.47

is

1.45

 हैं

1.44

Activations Density 0.539%