INDEX

Explanations

code data structures and types

The neuron fires on numeric literals in code or text (e.g. integers and floating‐point constants).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

Batman

0.55

Superman

0.52

Continental

0.52

Should

0.51

Telephone

0.49

Eastern

0.49

Jud

0.49

Arial

0.49

Barn

0.48

霑

0.48

POSITIVE LOGITS

 に対

0.56

ografia

0.54

 tiene

0.54

 cdot

0.52

 ที่

0.51

 datasets

0.50

ρ

0.49

 втра

0.49

[],

0.49

 chercher

0.49

Activations Density 0.000%