INDEX

Explanations

technical terms or commands

The neuron is sensitive to section headers and other structural boundaries in scholarly or technical documents (e.g. “Introduction,” “OBJECTIVES,” large numeric metadata blocks).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 אבל

1.16

but

1.14

 ஆனால்

1.14

 Nhưng

1.12

 nhưng

1.10

But

1.08

 কিন্তু

1.08

 pero

1.04

 ولكن

1.03

 Tetapi

1.03

POSITIVE LOGITS

Historically

0.86

近年来

0.83

Traditionally

0.82

Millions

0.76

 ലോക

0.71

Depuis

0.71

近年

0.70

Nowadays

0.68

越来越多的

0.68

 العديد

0.68

Activations Density 0.069%