INDEX

Explanations

mathematical or specialized terms

This neuron strongly activates on technical mathematical jargon—especially high-level terms like gradient, semigroup, maximal, operator, dissipation, etc.—in research-style abstracts.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

孕妇

-0.96

 Kassel

-0.81

Uniform

-0.81

Listing

-0.79

 groaning

-0.78

uniform

-0.78

Ĝ

-0.77

Investors

-0.75

 Vors

-0.75

 Finanz

-0.74

POSITIVE LOGITS

 Mather

0.79

 Unterkunft

0.77

віта

0.76

 От

0.75

 adquirir

0.73

spt

0.71

 ску

0.71

atecas

0.68

できるだけ

0.68

BV

0.68

Activations Density 0.027%