INDEX

Explanations

value of

The neuron sharply responds to the explicit “Start of …” and “End of …” section markers (often carrying large numeric values) that delimit different document segments.

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

 ہوسک

0.53

 любое

0.53

mid

0.52

 haters

0.49

 Vegas

0.49

phas

0.49

 midweek

0.48

телям

0.48

 کریں۔

0.48

গ্রন্থ

0.48

POSITIVE LOGITS

 Specification

0.68

 Assistance

0.63

vist

0.63

 Assists

0.58

 Initiative

0.57

Assist

0.57

vans

0.57

Zul

0.56

 Specifications

0.56

 Recordemos

0.55

Activations Density 0.000%