INDEX

Explanations

property key-value pairs

The neuron activates on numeric literals (multi-digit numbers) in the code.

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

 Wład

-1.27

Внеш

-1.27

 conster

-1.25

車は

-1.25

Однако

-1.23

他们

-1.23

Примеча

-1.23

鹟

-1.22

 eier

-1.22

Ké

-1.22

POSITIVE LOGITS

 from

1.55

liani

1.30

with

1.27

 verständ

1.23

 berücksich

1.21

if

1.20

for

1.19

ことも

1.16

 dokład

1.16

So

1.15

Activations Density 0.003%