INDEX

Explanations

starting statements or explanations

This neuron fires on first-person framing or transition phrases expressing the speaker’s intent or assumptions (e.g. “I will,” “I assume,” “I’ll”).

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

Ꮕ

-1.34

懈

-1.34

ianum

-1.22

pielen

-1.20

Zambia

-1.16

charCodeAt

-1.14

ragende

-1.13

蜮

-1.13

enada

-1.11

rechten

-1.11

POSITIVE LOGITS

first

1.66

 first

1.64

and

1.59

1.51

my

1.38

al

1.38

it

1.35

my

1.33

 ersten

1.29

ol

1.28

Activations Density 0.057%