INDEX

Explanations

geometric shapes and structures

The neuron activates on occurrences of the root “geometr-,” i.e. words like “geometric,” “geometrical,” or “geometrically.”

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

-2.34

-2.25



-2.20

?"

-2.17

-2.08

いなかった

-2.08

 tand

-2.05

 thai

-2.00

able

-1.99

袮

-1.94

POSITIVE LOGITS

 encantador

2.23

 欣赏

2.23

とは言

2.20

 estaremos

2.19

 größeren

2.09

 sentimos

2.02

 kurzen

2.02

2.00

鶏

1.97

 fantástico

1.96

Activations Density 0.013%