INDEX

Explanations

awards of the year

The neuron detects occurrences of “Year” in award-name phrases (e.g. “Officer of the Year,” “Coach of the Year,” etc.).

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

や

-1.82

过去

-1.70

因为

-1.66

 möjligt

-1.55

 одна

-1.55

\&

-1.54

дежда

-1.53

 только

-1.52

に

-1.47

假如

-1.46

POSITIVE LOGITS

for

1.84

 parle

1.82

 Canadá

1.73

 resté

1.70

 manuscrit

1.65

 facilite

1.64

 respekt

1.64

 fré

1.63

mlu

1.62

unción

1.61

Activations Density 0.012%