INDEX

Explanations

class definitions in code

The neuron is detecting Python definition keywords (e.g. “def” and “class”).

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

Its

-1.69

the

-1.68

 Their

-1.56

が無い

-1.38

hicles

-1.27

Of

-1.26

得罪

-1.25

nearly

-1.24

 myriad

-1.24

it

-1.24

POSITIVE LOGITS

1.60

خاص

1.51

<bos>

1.39

گزاری

1.35

我們

1.34

 evenement

1.33

Tá

1.31

нодоро

1.31

1.30

Activations Density 0.020%