INDEX

Explanations

if followed by pronouns

The neuron strongly activates on the conditional cue “if,” i.e. it detects occurrences of the word “if” marking hypothetical or conditional clauses.

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

被迫

-1.67

.”

-1.66

$-.

-1.62

,”

-1.60

psies

-1.60

);

-1.56

 $^{\

-1.54

”),

-1.54

之色

-1.54

⬮

-1.52

POSITIVE LOGITS

お楽しみに

1.70

 hanno

1.61

绚

1.59

return

1.50

 kazak

1.49

栳

1.48

的で

1.48

 inga

1.46

感じで

1.45

 lilla

1.45

Activations Density 0.011%