INDEX

Explanations

terms related to categories and data entities in a structured format

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

MigrationBuilder

-0.82

évaluateur

-0.81

WriteBarrier

-0.81

yarnpkg

-0.78

 виправивши

-0.78

 مشين

-0.77

 فريبيس

-0.77

 propOrder

-0.76

ंदीखरीदारी

-0.75

NameInMap

-0.75

POSITIVE LOGITS

0.54

’

0.45

0.43

is

0.39

0.38

may

0.36

 language

0.36

Music

0.36

Activations Density 0.255%

terms related to categories and data entities in a structured format

No Comments

No Known Activations