INDEX

Explanations

what are some

This neuron detects hedging or speculative language—non-specific quantifiers and modal terms (e.g. “some,” “possible,” “could,” “might,” “other,” etc.).

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

 holotype

-1.28

 during

-1.28

在此

-1.26

 each

-1.19

在中国

-1.18

紮

-1.17

閙

-1.16

在这个

-1.13

但是在

-1.11

一会

-1.10

POSITIVE LOGITS

 Lernen

1.51

PUTY

1.51

 discapacidad

1.50

 rempliss

1.47

she

1.46

of

1.45

が広

1.45

calça

1.40

：……

1.37

 Чтобы

1.34

Activations Density 0.013%