INDEX

Explanations

terms related to programming and marked code structures in software development

oai_token-act-pair · gpt-4o-mini Triggered by @bot

Mvx binding

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 betweenstory

-0.66

KURZBESCHREIBUNG

-0.63

 AttributeSet

-0.59

CloseOperation

-0.57

locaust

-0.56

 Bioaccumulative

-0.55

styleable

-0.55

AutoField

-0.54

 arşivlendi

-0.54

Portale

-0.54

POSITIVE LOGITS

caneca

0.41

DotNetBar

0.41

 Gestión

0.40

 remercie

0.40

 operadores

0.39

 cidad

0.39

 kertas

0.39

bricas

0.38

 hiszen

0.37

 cidade

0.36

Activations Density 0.001%

terms related to programming and marked code structures in software development

Mvx binding

No Comments

No Known Activations