INDEX

Explanations

acceptance

The neuron strongly fires on tokens that start a new sentence—that is, the very first word (or punctuation) immediately following a period.

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

 товар

0.79

 Telegraph

0.77

UART

0.77

 Эд

0.76

 себя

0.75

 выпол

0.75

CONTENT

0.74

続きを読む

0.74

 объект

0.73

 mitophagy

0.73

POSITIVE LOGITS

it

1.09

al

0.91

0.90

ono

0.87

ك

0.87

et

0.87

ar

0.86

াত্ত

0.85

iniai

0.85

urato

0.85

Activations Density 0.002%