INDEX

Explanations

equivalent to

The neuron fires on numeric data—especially floating‐point or large numerical values (statistics, metrics, counts) in the text.

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

chini

0.54

ATER

0.53

 ধাতু

0.52

➝

0.51

 দেবতা

0.50

 Mannes

0.50

apus

0.48

ISER

0.48

জ্জনক

0.48

POSITIVE LOGITS

multiple

0.59

更多的

0.56

 tỏa

0.56

ఎస్‌

0.55

飨

0.54

 múltipl

0.54

Lucky

0.53

 півні

0.53

persistence

0.52

 Fedora

0.52

Activations Density 0.001%