INDEX

Explanations

sorted lists

The neuron fires strongly on numeric tokens—especially floating‐point numbers with decimals.

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

不妨

0.84

عليه

0.76

Issuer

0.75

凳

0.75

Volatility

0.74

mdb

0.72

explain

0.72

謬

0.72

uild

0.71

Adjustment

0.70

POSITIVE LOGITS

 två

0.87

 lös

0.84

 först

0.82

 ৩৮

0.82

es

0.82

 fyra

0.80

ecz

0.79

ome

0.78

 epiz

0.78

그러나

0.77

Activations Density 0.000%