INDEX

Explanations

programming-related elements and structures

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

Gaussian kernel

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

 froh

-0.33

 asesinado

-0.32

 Unterscheidung

-0.32

 rechazo

-0.32

faßt

-0.32

chtigkeit

-0.30

 quirúrg

-0.30

 monasterio

-0.29

説明

-0.29

acakt

-0.29

POSITIVE LOGITS

 ProtoMessage

0.71

contentLoaded

0.71

تقاوى

0.69

enterOuterAlt

0.66

 OMITBAD

0.65

Personendaten

0.62

RotationOrder

0.61

 Савезне

0.60

səhifə

0.60

dafx

0.60

Activations Density 0.002%

programming-related elements and structures

Gaussian kernel

No Comments

No Known Activations