INDEX

Explanations

walking and gait

The neuron detects words and fragments related to walking, ambulation, and locomotion.

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

 hebben

-1.51

 både

-1.49

of

-1.45

 niitä

-1.43

 kedua

-1.31

 ARISING

-1.30

 bezüglich

-1.30

͢

-1.27

вають

-1.27

 вяза

-1.25

POSITIVE LOGITS

 Like

1.44

who

1.31

 Unlike

1.30

 amerikanischen

1.28

ка

1.28

 anglaise

1.27

but

1.27

and

1.23

超级

1.22

 Every

1.21

Activations Density 0.037%