INDEX

Explanations

greetings and exclamations

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

The neuron flags non-English or foreign tokens (especially Romanized Hindi/Indian words or other non-English script segments).

oai_token-act-pair · o4-mini Triggered by @jyhe0408

non-English (especially South Asian/Hindi-transliterated) words and names.

oai_token-act-pair · gpt-5 Triggered by @jyhe0408

Hindi or other Indic language text fragments (words or syllables).

oai_token-act-pair · claude-4-5-sonnet Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_10/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ーーーー

-2.19

焌

-2.19

ーーー

-2.09

Оценка

-2.08

Should

-2.05

有点

-1.99

というわけで

-1.96

真的很

-1.96

さてさて

-1.96

-1.91

POSITIVE LOGITS

 they

2.44

in

2.03

 that

1.91

铩

1.85

not

1.77

The

1.68

on

1.65

 itſelf

1.64

戆

1.63

😵

1.60

Activations Density 0.001%

greetings and exclamations

The neuron flags non-English or foreign tokens (especially Romanized Hindi/Indian words or other non-English script segments).

non-English (especially South Asian/Hindi-transliterated) words and names.

Hindi or other Indic language text fragments (words or syllables).

No Comments

No Known Activations

greetings and exclamations

The neuron flags non-English or foreign tokens (especially Romanized Hindi/Indian words or other non-English script segments).

non-English (especially South Asian/Hindi-transliterated) words and names.

Hindi or other Indic language text fragments (words or syllables).

No Comments

No Known Activations