INDEX

Explanations

This neuron strongly responds to temporal and numeric expressions—dates, times, counts, and related details.

then, later, afterward

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

 сахара

-1.41

 chré

-1.37

 چربی

-1.33

 géographie

-1.30

 religieuses

-1.26

 medan

-1.26

 αὐ

-1.25

 τὰ

-1.25

 összefoglal

-1.23

 ἐκ

-1.23

POSITIVE LOGITS

 then

2.25

 later

2.14

 subsequently

2.03

 subsequent

1.84

was

1.75

 when

1.73

ּוֹ

1.65

ַּ

1.55

ֶּ

1.51

後來

1.49

Activations Density 0.017%