INDEX

Explanations

economic observations and predictions

The neuron detects speculative or forecasting language—words that signal likelihood, prediction, or hedging.

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

 religieux

-2.05

diğim

-1.88

⁇

-1.87

adır

-1.85

ׇ

-1.81

 précieux

-1.75

 oscu

-1.74

㍑

-1.74

 dokter

-1.72

 保育

-1.72

POSITIVE LOGITS

now

2.39

ּוֹ

2.30

有些不

1.85

ִּ

1.77

 will

1.73

styrelsen

1.73

 موبایل

1.62

are

1.62

 traducción

1.61

ֶּ

1.60

Activations Density 0.235%