INDEX

Explanations

dates 2023 / early 2024

The neuron activates on numeric tokens (numbers, dates, and numeric sequences).

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

?).

0.57

lardan

0.54

 réserve

0.54

zte

0.53

全年

0.53

thin

0.52

lara

0.52

로나

0.52

 akkor

0.52

 கலோர

0.50

POSITIVE LOGITS

밤

0.61

၁

0.59

SHE

0.57

 Bist

0.57

 Penal

0.56

AnyCPU

0.56

 Emit

0.55

 Bhajan

0.55

 lading

0.55

 இரவு

0.54

Activations Density 0.083%