INDEX

Explanations

patterns related to structured data formats

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

 Administrativna

-0.67

 Italijanski

-0.65

 AssemblyProduct

-0.61

saraba

-0.58

 OFDb

-0.57

URBANA

-0.55

Autoritní

-0.55

findpost

-0.52

 betweenstory

-0.52

Datuak

-0.51

POSITIVE LOGITS

mapStateToProps

0.38

 potentially

0.38

 potential

0.38

 this

0.36

 within

0.34

 contigo

0.34

گاهی

0.33

 condena

0.33

اریخ

0.33

 conmigo

0.32

Activations Density 0.316%

patterns related to structured data formats

No Comments

No Known Activations