INDEX

Explanations

physics and motion concepts

The neuron fires on tokens naming physical motion parameters in code (e.g. speed, velocity, force, acceleration).

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

lgica

-1.34

mentar

-1.34

dekat

-1.33

냑

-1.23

yczny

-1.23

 одежда

-1.23

¹,

-1.22

nimo

-1.20

komme

-1.20



-1.20

POSITIVE LOGITS

<bos>

1.53

我跟

1.30

 tendremos

1.28

1.26

俢

1.23

1.22

帮我

1.20

 podremos

1.20

 ermittelt

1.20

 女人

1.16

Activations Density 0.009%