INDEX

Explanations

different languages for "I"

This neuron detects the assistant's self-referential first-person statements (claims about its abilities, preferences, or actions).

New Auto-Interp

Configuration

Prompts (Dashboard)

238,145 prompts, 512 tokens each

Dataset (Dashboard)

lmsys + oasst1

Embeds

IFrame

Link

Negative Logits

 bize

0.82

 ourselves

0.77

给我们

0.76

讓我們

0.75

 وقلنا

0.73

 nám

0.71

 нам

0.70

 vimos

0.70

让我们

0.69

 вам

0.69

POSITIVE LOGITS

 আমি

2.05

 நான்

2.05

 मैं

1.98

私は

1.98

 నేను

1.93

 tôi

1.84

我会

1.83

 ನಾನು

1.82

 ഞാൻ

1.81

 मैंने

1.79

Activations Density 0.155%