INDEX

Explanations

elements that indicate class structures and definitions in code-related contexts

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 فريبيس

-0.82

 صوتيه

-0.81

 '\\;'

-0.72

ViewFeatures

-0.71

UrlResolution

-0.71

SharedDtor

-0.71

 estekak

-0.70

 <<<<<<<<<<<<<<

-0.69

BeginInit

-0.67

Diweddarwch

-0.66

POSITIVE LOGITS

iapkan

0.39

 suivant

0.33

 הי

0.32

 Herz

0.31

map

0.30

joseph

0.29

姉妹

0.29

many

0.29

haben

0.28

 Abfall

0.28

Activations Density 0.437%

elements that indicate class structures and definitions in code-related contexts

No Comments

No Known Activations