INDEX

Explanations

calculate and compute values

This neuron detects the main action word at the start of math or data‐analysis question headings—especially verbs like “Computing,” “Calculating,” or “Calculation.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 négociations

-1.19

 enthou

-1.04

 aident

-1.01

鐽

-1.00

@"/

-0.99

ቕ

-0.98

不喜欢

-0.98

せっかく

-0.96

 tedes

-0.94

measured

-0.91

POSITIVE LOGITS

how

1.85

 from

1.52

 required

1.45

 using

1.42

 whether

1.40

 needed

1.32

 what

1.27

 necessary

1.27

 appropriate

1.25

for

1.23

Activations Density 0.090%