INDEX

Explanations

you and we referring to groups

The neuron strongly activates on second-person pronouns, especially “you,” marking direct addresses to the reader.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 have

0.88

have

0.76

 Have

0.75

Have

0.71

 HAVE

0.70

 aren

0.66

 haven

0.61

are

0.61

 tienes

0.59

 داریم

0.58

POSITIVE LOGITS

 guys

1.55

 ребята

1.24

guys

1.20

 lads

1.13

 muchachos

1.13

 chercheurs

1.13

 bloggers

1.09

 ragazzi

1.06

 artisans

1.06

 meninos

1.06

Activations Density 0.233%