INDEX

Explanations

comparative adverbs

This neuron detects Russian words in comparative or superlative form (e.g. those ending in –ее, –е, or comparable adjectives/adverbs like “хуже,” “неприятнее,” “менее,” “точнее”).

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

in

-2.97

</strong>

-2.70

as

-2.45

it

-2.41

at

-2.39

or

-2.19

for

-2.09

咫

-2.06

酩

-2.02

岌

-1.95

POSITIVE LOGITS

2.73

 toalha

2.55

 démocr

2.47

 jahre

2.27

what

2.23

 noires

2.19

他的

2.16

 diriger

2.14

也沒有

2.14

It

2.11

Activations Density 0.004%