INDEX

Explanations

discrepancy between or among

The neuron activates on mentions of “discrepancy” (or “discrepancies”) indicating differences or mismatches in measurements or values.

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

 were

-2.95

You

-2.69

你有

-2.58

 テープ

-2.53

];

-2.47

 ваше

-2.47

derdag

-2.45

 или

-2.31

 Palabra

-2.27

 Your

-2.25

POSITIVE LOGITS

2.84

↵

2.67

up

2.42

</h3>

2.30

ついでに

2.23

ika

2.20

鈇

2.13

じゃあ

2.05

 regal

2.02

Activations Density 0.005%