INDEX

Explanations

'Xe' followed by text fragments

The neuron detects occurrences of the Greek lowercase letter “α.”

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

-2.98

-2.91

isation

-2.27

驁

-2.27

 säger

-2.25

ًا

-2.22

诸多

-2.17

 decreased

-2.08

 flesta

-2.03

 angebra

-2.02

POSITIVE LOGITS

be

3.80

 μην

2.55

2.39

 δύο

2.22

翃

2.13



2.09

 современной

2.08



2.06

珢

2.06

Activations Density 0.000%