INDEX

Explanations

numeric values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

JspWriter

-0.52

->___

-0.51

oare

-0.46

WebElementEntity

-0.45

UAWEI

-0.44

DIPSETTING

-0.44

ItemBackground

-0.44

 Europas

-0.43

 uLocal

-0.43

社の

-0.42

POSITIVE LOGITS

出版年

0.57

httphttps

0.47

rungsseite

0.43

 viewDidLoad

0.41

 CreateTagHelper

0.41

Diwedd

0.39

 nakalista

0.38

 insuffisamment

0.37

 ligiloj

0.36

таратура

0.36

Activations Density 0.223%

numeric values

No Comments

No Known Activations