INDEX

Explanations

import and use statements

This neuron strongly activates on structural/formatting tokens—especially the '<' symbol and similar markup or punctuation delimiters.

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

ว

0.71

τηση

0.66

匚

0.66

Ԁ

0.64

 proposent

0.63

efeated

0.61

ৃ

0.61

LaunchScheme

0.60

пени

0.60

 deprecated

0.60

POSITIVE LOGITS

 prenda

0.65

 irregularity

0.62

Others

0.62

आधी

0.62

सर्व

0.61

扱い

0.61

Berger

0.61

Vé

0.60

埃

0.60

 emboss

0.59

Activations Density 0.000%