INDEX

Explanations

context of numeric quantities

The neuron is tuned to pick out scientific reporting language that signals quantitative measurements or changes (e.g. words like increases, reduction, variability, standard deviation, elevated, measured).

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

 APPELLANT

-1.19

 conçu

-1.13

 rebajas

-1.11

 книге

-1.10

 manguera

-1.07

 ativar

-1.06

瘜

-1.06

($.

-1.06

 travailler

-1.02

 evaluaciones

-1.01

POSITIVE LOGITS

 overall

1.06

这里的

0.92

for

0.91

 because

0.90

 avere

0.90

傀

0.88

靭

0.87

in

0.85

لاف

0.85

 sehr

0.84

Activations Density 0.216%