INDEX

Explanations

by + [default, topic, car, proximity, their, the, others, size]

The neuron strongly activates on occurrences of “By” when it’s used to introduce a date or time reference.

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

ক

1.70

น้อง

1.57

ש

1.56

的所有

1.38

기

1.32

ü

1.30

 пройдет

1.29

ной

1.27

stricken

1.27

গত

1.23

POSITIVE LOGITS

zantine

2.34

 virtue

2.23

 dint

1.81

 means

1.52

 mistake

1.45

 default

1.39

zant

1.38

 implication

1.38

 separado

1.35

 analogy

1.29

Activations Density 0.435%