INDEX

Explanations

questions with parenthetical options

This neuron fires on question marks, flagging interrogative sentences.

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

 nhưng

0.91

こちらは

0.89

但也

0.86

 όλ

0.81

 Nhưng

0.80

 lakini

0.80

 Şimdi

0.79

 années

0.79

BUT

0.79

lında

0.77

POSITIVE LOGITS

?"

1.70

?”

1.52

？」

1.37

]?

1.30

?)

1.30

()?

1.27

？

1.25

؟

1.24

?\

1.23

?.

1.21

Activations Density 0.164%