INDEX

Explanations

on

The neuron detects words that introduce instructions or information (e.g. “instructions,” “details,” “information,” “how to”).

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

!,

0.72

함으로써

0.70

!',

0.66

]);

0.64

 형태로

0.61

 brainchild

0.61

]),

0.61

 indirectly

0.61

!),

0.61

하거나

0.59

POSITIVE LOGITS

详

0.89

详情

0.81

詳

0.80

說明

0.77

ガイド

0.75

 안내

0.75

 guides

0.73

说明

0.73

 подробно

0.73

 GUIDE

0.73

Activations Density 0.090%