INDEX

Explanations

limits and conditions

The neuron is keyed to common function words—especially pronouns (e.g. “they”), auxiliaries and modals (e.g. “can,” “have”), and relative words (e.g. “which”).

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

ablen

-1.35

Kör

-1.32

 оригіналу

-1.27

atonin

-1.23

ETHING

-1.23

orance

-1.23

štění

-1.19

llte

-1.18

eavor

-1.17

碚

-1.16

POSITIVE LOGITS

you

1.80

may

1.64

1.52

on

1.47

 when

1.38

 может

1.30

 might

1.30

 people

1.28

1.27

1.24

Activations Density 0.127%