INDEX

Explanations

The neuron fires on comparative or quantitative intensity markers—words and phrases that express greater or lesser magnitude (e.g. “times,” “stronger,” “larger,” “slower,” “higher,” “lower,” etc.).

comparative adjectives and adverbs

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

 cómoda

-1.25

ꈊ

-1.11

 écoute

-1.09

 acoged

-1.08

 SUCH

-1.05

こない

-1.02

 regroupe

-1.02

 cómodo

-1.02

 alapján

-1.02

 никогда

-1.01

POSITIVE LOGITS

 than

3.28

of

2.06

 compared

1.84

to

1.81

 from

1.55

 when

1.55

as

1.44

 rispetto

1.30

istani

1.20

 sedangkan

1.20

Activations Density 0.060%