INDEX

Explanations

programming-related terms and structures, specifically classes and models

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

domain driven models

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

Negative Logits

 wiser

-0.41

 poudre

-0.39

 NACIONAL

-0.39

 rodea

-0.39

をかけて

-0.38

 Giri

-0.38

 estekak

-0.38

riwal

-0.38

 Kriege

-0.38

ainya

-0.37

POSITIVE LOGITS

ViewFeatures

0.63

NameInMap

0.58

msgTypes

0.57

oredCriteria

0.54

InitVars

0.46

DataPropertyName

0.45

Становништво

0.45

JoinColumn

0.45

GEBURTS

0.44

EndInit

0.43

Activations Density 0.005%

programming-related terms and structures, specifically classes and models

domain driven models

No Known Activations