INDEX

Explanations

error codes and system call-related terms

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

adaptiveStyles

-0.42

UseVisualStyle

-0.40

ngdoc

-0.39

addCriterion

-0.33

Toponymie

-0.33

insatz

-0.33

 Мексичка

-0.32

okovic

-0.32

upassen

-0.32

Autoritní

-0.30

POSITIVE LOGITS

setupUi

0.57

AndEndTag

0.56

Tembelea

0.53

bers

0.48

 altså

0.47

 nakalista

0.47

 pallets

0.45

ScopeManager

0.45

 popoli

0.44

 iprot

0.44

Activations Density 0.045%

error codes and system call-related terms

No Comments

No Known Activations