INDEX

Explanations

C/C++ and VBA code keywords

The neuron activates on numeric literals (digits or 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

pso

-0.92

ETHODS

-0.91

 グラフ

-0.90

akteur

-0.88

 meanwhile

-0.87

колко

-0.87

 incluyen

-0.86

RTSN

-0.85

あとは

-0.85

その他の

-0.85

POSITIVE LOGITS

～」

0.90

zten

0.89

 again

0.88

----+

0.85

 Hache

0.84

ysaccharide

0.84

 they

0.84

ادا

0.83

TRACE

0.83

 placa

0.83

Activations Density 0.015%