INDEX

Explanations

achievements and roles

The neuron fires on mentions of sports game outcomes, especially numerical scores and “win/victory” contexts.

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

 funktioniert

0.61

 jeden

0.58

തല്ല

0.55

 შემთხვევაში

0.54

 unsere

0.53

 unspoken

0.52

 diejenigen

0.52

 স্পষ্ট

0.51

 হয়ত

0.51

 součást

0.50

POSITIVE LOGITS

 στο

0.84

 στην

0.74

 tại

0.70

 στη

0.68

ở

0.67

în

0.67

 nella

0.66

on

0.64

在大

0.64

 downlink

0.64

Activations Density 0.006%