INDEX

Explanations

each of parameterization

The neuron fires on tokens involved in subscripted mathematical identifiers (e.g. “I_”, “n_”, “s_”, array or map indexing).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 dör

-0.90

राब

-0.89

 própria

-0.87

スタル

-0.86

 humanas

-0.85

 kred

-0.84

reams

-0.82

 már

-0.80

 jaja

-0.80

 túnel

-0.79

POSITIVE LOGITS

 each

1.34

 jeweiligen

0.98

 каждого

0.92

每個

0.89

 bawat

0.88

each

0.88

 mỗi

0.87

 Each

0.83

Each

0.82

 corresponding

0.82

Activations Density 0.035%