INDEX

Explanations

The neuron fires on auxiliary/modal verbs (e.g. “would,” “should”) that mark polite requests, suggestions, or conditional mood.

noun followed by another noun

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

 from

-1.25

to

-1.24

 that

-1.20

-1.16

xxi

-1.15

ทุก

-1.12

 sebelum

-1.12

kraine

-1.10

 allem

-1.09

in

-1.09

POSITIVE LOGITS

是怎么回事

1.28

Handlung

1.15

市的

1.12

BIOL

1.08

̀u

1.05

/////////////

1.05

oherty

1.05

并未

1.04

锒

1.03

//{

1.02

Activations Density 0.101%