INDEX

Explanations

test context

The neuron fires on occurrences of “test” in code identifiers—especially in function or class names for unit tests (e.g. tokens like “test_” or “Test…”).

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

ſtre

-1.10

mentaux

-1.10

テスト

-1.08

測試

-1.06

솝

-1.05

ዴ

-1.05

趼

-1.05

鸩

-1.04

Ouverture

-1.03

torem

-1.01

POSITIVE LOGITS

 that

2.08

 getting

1.47

 basic

1.46

get

1.45

for

1.37

 something

1.35

 correct

1.33

 various

1.31

if

1.30

 against

1.29

Activations Density 0.017%