INDEX

Explanations

should not or be

The neuron activates strongly on occurrences of the modal verb “should,” signaling recommendations or normative advice.

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

ó

1.46

ದ್

1.35

 exhort

1.35

та

1.34

 imprison

1.27

д

1.27

 phá

1.27

со

1.25

ą

1.23

SCORE

1.20

POSITIVE LOGITS

有的

1.76

ی

1.75

ろん

1.54

alten

1.52

ात

1.51

ところ

1.50

 rechten

1.45

 helst

1.45

eh

1.44

 universitaria

1.43

Activations Density 0.144%