INDEX

Explanations

sample or example

The neuron highlights salient content words—mainly nouns and verbs carrying the core meaning of a sentence—while ignoring function words.

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

 meski

0.86

iyo

0.81

 nuanced

0.80

 🎉

0.79

 crafting

0.78

✨

0.77

 lätt

0.77

 nestled

0.77

 สร้าง

0.77

 Dzięki

0.77

POSITIVE LOGITS

但是在

0.76

validation

0.75

intention

0.71

に関して

0.67

scheduler

0.66

っいて

0.66

opinion

0.66

 solamente

0.66

しか

0.65

previous

0.65

Activations Density 0.001%