INDEX

Explanations

code punctuation and structure

The neuron activates on standalone numeric literals (multi‐digit numbers) in 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

bote

-1.12

/).

-1.04

:"-

-1.02

öff

-1.02

thd

-0.98

\}.

-0.98

 veía

-0.97

certification

-0.94

seye

-0.94

 practicar

-0.94

POSITIVE LOGITS

 DURING

1.18

發表

1.08

 古い

1.08

さあ

1.05

ből

1.05

を表示

1.05

＊＊＊＊

1.04

Gebruik

1.02

點擊

1.02

 magister

1.00

Activations Density 0.005%