INDEX

Explanations

revealing new truths

The main thing this neuron does is pick out factual report language—especially named entities, dates, numeric data, and revelations of new information.

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

on

-0.98

 experiences

-0.91

≾

-0.89

ensors

-0.88

for

-0.82

 here

-0.82

 cars

-0.82

涅

-0.81

현

-0.80

joy

-0.79

POSITIVE LOGITS

 revelation

2.16

 revealed

2.14

 revelations

1.94

 reveal

1.84

 discovered

1.77

原來

1.73

原来

1.73

 reveals

1.66

 discovery

1.65

 Revealed

1.56

Activations Density 0.081%