INDEX

Explanations

incident occurrence timing

The neuron activates most strongly on reporting verbs and phrases that introduce when or how an event happened (e.g. “incident occurred,” “took place,” “happened”).

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

幛

-1.70

妡

-1.61

ierten

-1.48

ㆍ

-1.47

𓆜

-1.45

mment

-1.44

brium

-1.41

signable

-1.40

ॐ

-1.40

ingt

-1.39

POSITIVE LOGITS

↵

1.66

1.55

 đã

1.48

 sudah

1.34

dengan

1.31

りました

1.30

Another

1.30

 unseren

1.28

if

1.27

もありました

1.27

Activations Density 0.019%