INDEX

Explanations

uniform distribution and random numbers

The neuron strongly activates on numeric tokens and numeric‐type identifiers in code (e.g. number literals like “0” or “1” and types like “double,” “float,” “int”).

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

ỗi

-0.77

 GUEST

-0.76

ropol

-0.76

Black

-0.76

 edema

-0.73

 рассчиты

-0.73

 plagioclase

-0.72

orced

-0.72

 биз

-0.72

Vey

-0.72

POSITIVE LOGITS

 uniform

1.53

 Uniform

1.38

uniform

1.37

Uniform

1.26

 uniformes

1.20

 uniforms

1.16

 rand

1.12

 uniformly

1.04

 unif

1.03

rand

0.98

Activations Density 0.018%