INDEX

Explanations

under conditions or supervision

This neuron fires on occurrences of the word “under” when it introduces a numerical value (e.g. “under 2012,” “Under 2143.95,” etc.).

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

ب

1.45

ልክ

1.36

1.34

vgl

1.26

1.22

你

1.18

raster

1.17

৮

1.16

yorum

1.16

anı

1.15

POSITIVE LOGITS

 auspices

1.98

neath

1.77

 purview

1.55

dogs

1.52

𝘽

1.42

estimating

1.38

 guise

1.23

 Under

1.21

 circumstances

1.21

𝘾

1.20

Activations Density 0.043%