INDEX

Explanations

programming, symbols, or multilingual words

This neuron fires on numeric tokens and mathematical notation (e.g. numbers, arrows, and other math‐related symbols).

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

They

-2.77

ritsar

-2.66

』

-2.53

 этими

-2.44

 azienda

-2.42

There

-2.36

她

-2.34

࿙

-2.34

//"

-2.33

鴣

-2.33

POSITIVE LOGITS

’

2.91

 diejenigen

2.61

2.55

なんだか

2.47

2.45

椟

2.44

 necessariamente

2.38

“

2.38

2.36

Activations Density 0.001%