INDEX

Explanations

relative pronouns or connectors

The neuron strongly activates on longer, content‐heavy words (especially multi‐syllable nouns and technical terms).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 significantly

0.70

 அவர்க

0.64

தின

0.60

 কোন

0.57

प्रिय

0.55

 आम्ही

0.55

 असल्या

0.55

 kanya

0.54

ленными

0.53

之

0.52

POSITIVE LOGITS

 that

3.66

that

3.52

 που

3.23

 که

3.17

që

2.92

 التي

2.83

 THAT

2.59

 thats

2.58

THAT

2.53

 которые

2.48

Activations Density 2.253%