INDEX

Explanations

stock trading and prices

The neuron is primarily detecting occurrences of the word “stock” (often as part of compound or hyphen/underscore‐joined technical terms like “stock‐keeping unit,” “stock_figures,” or “stock‐removed”).

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

imetsu

-1.80

-1.74

 souhait

-1.71

-1.61

–

-1.58

ce

-1.58

ت

-1.55

}$

-1.53

 Tonnen

-1.53

сля

-1.51

POSITIVE LOGITS

 motivadoras

1.79

zzlement

1.77

 pertumbuhan

1.77

几人

1.75

 ткани

1.75

</em>

1.74

 Stock

1.74

_));

1.73

 магазине

1.72

 inspiradoras

1.72

Activations Density 0.009%