INDEX

Explanations

pairwise calculations between items

The neuron strongly activates on names of graph‐theoretic or combinatorial optimization problems (e.g. “traveling salesman,” “pairwise,” “adjacency matrix”).

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

nef

-1.20

izieren

-1.15

र्शन

-1.13

↵↵

-1.12

Puede

-1.09

fig

-1.09

ᚕ

-1.06

により

-1.05

薬品

-1.05

‚¬

-1.03

POSITIVE LOGITS

 folgende

1.41

poň

1.23

 doká

1.21

 będzie

1.17

 between

1.16

繞

1.13

 こう

1.13

 deberán

1.13

ované

1.13

 these

1.09

Activations Density 0.016%