INDEX

Explanations

activate windows 7 installation

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

The neuron strongly activates on numeric tokens (digits and number sequences), i.e. it finds standalone numbers in the text.

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

references to past time periods, especially specific years or decades.

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

New Auto-Interp

Configuration

google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ఇప్పటికే

0.80

 består

0.76

 विशेषज्ञों

0.76

 gerekiyor

0.72

 ஏற்பட்டு

0.71

 আপাতত

0.71

 జరుగు

0.68

ऑनलाइन

0.67

 prochains

0.67

 seit

0.66

POSITIVE LOGITS

当時の

1.86

当時

1.85

那时候

1.84

 damals

1.83

 시절

1.81

當時

1.79

 당시

1.77

那时

1.71

当时的

1.67

当时

1.63

Activations Density 1.254%

activate windows 7 installation

The neuron strongly activates on numeric tokens (digits and number sequences), i.e. it finds standalone numbers in the text.

references to past time periods, especially specific years or decades.

No Comments

No Known Activations

activate windows 7 installation

The neuron strongly activates on numeric tokens (digits and number sequences), i.e. it finds standalone numbers in the text.

references to past time periods, especially specific years or decades.

No Comments

No Known Activations