INDEX

Explanations

phrases and terms associated with artistic or entertainment achievements and careers

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

ggak

-0.30

【

-0.29

ätte

-0.27

bezpiec

-0.27

 añad

-0.27

 offered

-0.26

 werd

-0.26

 Auß

-0.26

Easy

-0.25

———

-0.25

POSITIVE LOGITS

 AssemblyCompany

0.93

CppMethod

0.77

AddTagHelper

0.74

ViewImports

0.70

richTextPanel

0.69

AndEndTag

0.65

uxxxx

0.63

LookAnd

0.62

GraphicsUnit

0.61

ValueStyle

0.61

Activations Density 0.025%

phrases and terms associated with artistic or entertainment achievements and careers

No Comments

No Known Activations