INDEX

Explanations

over a number of years

The neuron detects multi‐digit numeric tokens representing large numbers (e.g. years or big counts).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 fifty

0.88

 sixty

0.84

 seventy

0.80

 forty

0.77

fifty

0.72

 eighty

0.71

 fift

0.66

 ninety

0.65

 half

0.64

 Fifty

0.63

POSITIVE LOGITS

dB

0.44

Rq

0.44

abh

0.43

یدر

0.42

 중에

0.41

abl

0.41

 없고

0.41

Insertion

0.41

ω

0.41

dB

0.40

Activations Density 0.107%