INDEX

Explanations

time of day

The neuron fires on words denoting times of day (e.g. morning, evening, night) in context.

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

下午

1.27

上午

1.25

 afternoon

1.23

 morning

1.15

 صباح

1.14

早上

1.13

 오전

1.13

中午

1.11

 दोपहर

1.10

 দুপুর

1.08

POSITIVE LOGITS

 snack

0.72

 Agar

0.69

 ವೇಳೆ

0.66

 Iterable

0.65

 snacks

0.63

clubs

0.61

 taxis

0.59

 trains

0.58

 অ্যাস

0.58

io

0.57

Activations Density 0.173%