INDEX

Explanations

but to, we maintain

The neuron primarily detects numerical tokens—especially statistics or measure-like numbers—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

 beheld

0.81

 expropri

0.77

 etsy

0.76

 quaint

0.75

 feminist

0.74

 widowed

0.74

 insolvent

0.73

 layer

0.72

 gently

0.72

 pyrolysis

0.72

POSITIVE LOGITS

 선수

1.51

球员

1.48

球队

1.44

 Coaches

1.43

 teammates

1.41

 بازیکن

1.41

選手

1.37

 jogadores

1.37

 тренер

1.35

 Mannschaft

1.34

Activations Density 0.548%