INDEX

Explanations

self-reference ("I")

np_acts-logits-general · gemini-2.5-flash-lite

first-person self-reference pronouns, my, me

np_acts-logits-general · gemini-2.5-flash-lite

say "I"

np_max-act-logits · claude-4-5-sonnet

Triggered by @sripadkarne

New Auto-Interp

Configuration

google/gemma-scope-2-27b-pt/resid_post/layer_31_width_16k_l0_medium

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

لة

0.40

స్తారు

0.40

lucent

0.40

ល់

0.39

')",

0.39

นาะ

0.38

esModule

0.37

жээ

0.37

the

0.37

كله

0.36

POSITIVE LOGITS

 আমি

1.84

我的

1.83

 tôi

1.83

我

1.83

 моей

1.81

 мне

1.80

 मैंने

1.79

 ನಾನು

1.79

 আমিও

1.77

 मैं

1.76

Activations Density 1.707%

self-reference ("I")

first-person self-reference pronouns, my, me

say "I"

No Known Activations

self-reference ("I")

first-person self-reference pronouns, my, me

say "I"

No Known Activations