INDEX

Explanations

table borders and formatting

The neuron spikes on long runs of punctuation characters used as ASCII-art borders or separators (e.g. the repeated “–”, “+”, “═”, etc. that draw table/grid lines).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 dibutuhkan

-1.25

 diperlukan

-1.23

-1.20

chaise

-1.20

bluza

-1.20

++)

-1.19

 Entscheidungen

-1.19

︠

-1.16

 gerekiyor

-1.16

 either

-1.13

POSITIVE LOGITS

却不

1.16

1.11

</em>

1.09

1.06

Ine

1.06

<td>

1.05

 résulte

1.04

1.03

 अच्छ

1.02

Activations Density 0.001%