INDEX

Explanations

we are, we want

The neuron responds strongly to personal pronouns—especially first‐person “I” and “we” (and to a lesser extent “you”)—marking speaker references.

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

MEDIUM

0.60

BinaryOperation

0.59

៣

0.59

JScripts

0.57

保証

0.57

 อ่า

0.57

ヽ

0.56

小さ

0.55

rscheinlich

0.55

Fragments

0.55

POSITIVE LOGITS

 while

0.77

 along

0.73

 While

0.68

 whilst

0.67

 however

0.62

㭐

0.62

 Along

0.61

 malgré

0.61

 endeavour

0.60

 though

0.59

Activations Density 0.157%