INDEX

Explanations

morning times (AM/PM)

This neuron identifies the numeric tokens that specify times (hours or minutes) in time expressions (e.g., “9:57 PM,” “4:03 AM,” minute counts) in questions or statements about time.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

-1.55

<i>

-1.35

 December

-1.34

();

-1.31

 about

-1.30

 November

-1.27

 October

-1.26

</h2>

-1.26

tbe

-1.23

 September

-1.20

POSITIVE LOGITS

of

1.32

 Teknologi

1.27

 idee

1.26

睡眠

1.25

bebasan

1.24

 naší

1.23

 teknologi

1.20

凌晨

1.17

ussure

1.14

 principes

1.13

Activations Density 0.059%