INDEX

Explanations

see what

The neuron spikes on verbs of perception—especially instances of “see” (and similar perceiving verbs).

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

 trasport

1.63

𝐚

1.63

 sepen

1.56

ی

1.55

metik

1.48

𝐄

1.47

 diretamente

1.42

ICAGO

1.41

𝐨

1.41

 özellik

1.41

POSITIVE LOGITS

 glimpses

1.52

ות

1.43

к

1.38

 firsthand

1.23

टर

1.20

ੜ

1.19

त

1.18

ле

1.16

 plight

1.16

тю

1.16

Activations Density 0.169%