INDEX

Explanations

decimal places or units

The neuron strongly activates on four‐digit numbers in the 1900s (i.e. year/date tokens).

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

-1.90

<bos>

-1.78

ناسب

-1.77

 accusing

-1.73

 упраж

-1.62

 including

-1.61

 które

-1.59

改为

-1.59

剩下

-1.58

ꠥ

-1.58

POSITIVE LOGITS

 mijne

2.19

 zoude

2.17

鹉

1.97

 "...

1.86

 ingenting

1.84

噠

1.84

 droom

1.81

桖

1.80

</h6>

1.80

$};

1.80

Activations Density 0.035%