INDEX

Explanations

terms related to powered and electric vehicle features

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

automatic power features

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



-0.56

TypedDataSet

-0.50

 تضيفلها

-0.50

፩

-0.40

Архівовано

-0.39

 neod

-0.39

 gewor

-0.38

addCell

-0.38

KommentareTeilen

-0.38

 validity

-0.38

POSITIVE LOGITS

parsedMessage

0.65

 automatico

0.58

automatic

0.57

AnimationsModule

0.56

Automatic

0.52

automated

0.50

 automati

0.50

 automatic

0.49

featureID

0.48

QMetaType

0.48

Activations Density 0.010%

terms related to powered and electric vehicle features

automatic power features

No Known Activations