INDEX

Explanations

non-consensual acts

This neuron responds to hyphen or dash tokens used in technical or compound constructions (e.g. “non-,” “pre-,” “post-”) rather than ordinary text.

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

0.92

’-

0.71

esque

0.61

stä

0.59

0.57

'-

0.55

 whereupon

0.55

}-

0.55

Ў

0.55

making

0.54

POSITIVE LOGITS

ニス

0.54

ุ้น

0.48

 परेशानी

0.47

KERNEL

0.47

नहीं

0.47

вами

0.46

quorum

0.46

<unused2164>

0.45

ニック

0.45

мый

0.45

Activations Density 0.218%