INDEX

Explanations

choice or frequency

The neuron strongly activates on formal, technical descriptions and definitions of processes, devices, or compounds—i.e. the patent‐style/specification language sections.

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

 when

-1.86

 this

-1.73

 there

-1.59

it

-1.54

 those

-1.49

 just

-1.44

 during

-1.43

 only

-1.32

now

-1.23

 immediately

-1.23

POSITIVE LOGITS

 either

3.58

either

3.48

Either

3.44

 либо

3.14

 entweder

3.05

usually

2.81

 Either

2.81

 usually

2.75

 généralement

2.52

 или

2.47

Activations Density 0.364%